In a survey form, which options would be the best (user friendly)?

1. Are you male or female? 
o Female 
o Male

2. Are you female or male? 
o Female 
o Male

  • 59
    "Gender" as title. Not sure why do you want to stress "Are you....".
    – Amit Jain
    Commented Mar 28, 2017 at 9:46
  • 63
    And those who don't identify themselves as either one, what should they mark? Commented Mar 28, 2017 at 12:47
  • 82
    @BennySkogberg The form is clearly not asking "Do you identify as..."
    – MonkeyZeus
    Commented Mar 28, 2017 at 13:18
  • 16
    So the short answer to your original question, @Pradeep, is apparently a resounding "neither." Commented Mar 28, 2017 at 13:48
  • 18
    @MonkeyZeus I saw a question on here recently that asked something like: "what if my birth country no longer exists?" What indeed. Why is nationality so darned important? Age? Height? Weight? This is the Internet, no one knows you are a dog. (and in many cases it is true)
    – user67695
    Commented Mar 28, 2017 at 14:06

15 Answers 15


tl;dr: No matter what you do, you will be in good company. Purely as far as order goes male-first seems to be the traditional preference and seems to match linguistic expectations of most users. Despite this modern western tech companies seem to opt for female-first.

Cultural connotations

A lot of the answers dive into western liberal notions about sexuality and gender and the perception of the question itself. Although this is an interesting and fascinating philosophical question it's hardly relevant for non-western cultures where asking for sex is still a normal question. This can be normal because of grammatical reasons (addressing a person without knowing their gender is not possible in some languages (or at the very least most indo european languages don't allow you to talk about a person without knowing their gender)) or simply because a lot of cultures embrace different gender roles.

Especially if you are addressing a purely western audience it is however a good idea to consider whether you need this information. From a UX perspective decreasing friction by not including the question (or making it optional) is advantageous, however it comes at its own cost.

So, back to the actual question

Usage across texts and the entire internet

Without a particularly strong UX reason to deviate from the norm the best rule of thumb is to match the expectations the user will have.

One way to get an impression of user expectations is by looking at Google ngrams which shows term usage in books (only including a minimal amount of forms of course):

enter image description here

Another option is looking at google search results: 21,100,000 vs 570,000 results respectively for "male or female" or "female or male".

I checked the same for Dutch, German and Slovak and this seems to hold true across a variety of western cultures, so from that point of view mentioning 'male' first seems to be the best choice.

Usage exclusively on registration forms on most popular websites

Secondly I wanted to take a look at what most popular (as defined by the Alexa site ranking) websites use:

  • Google: non-binary dropdown form with 'female' first
  • Youtube: see Google
  • Facebook: binary with 'female' first
  • Baidu: 'male' first
  • Wikipedia: doesn't ask
  • Yahoo: free form input with suggested option list with 'female' first
  • Google India: see Google
  • Amazon: doesn't ask
  • Tencent QQ: 'male' first
  • Google Japan: see Google
  • Windows Live: non-binary dropdown, 'male' first
  • Taobao: Need Chinese phone number to get to 'personal information' part of the form
  • VK.com: 'female' first
  • Twitter: Doesn't ask
  • Instagram: Doesn't ask
  • Hao123: See baidu
  • Sohu: Doesn't ask


My conclusion from this is that: 1) Traditionally 'male' first had a strong preference 2) Silicon Valley companies seem to strongly prefer 'female'-first.

Either way: You will be fine whatever you pick and considering the diversity of orders and inputs types on popular registration forms it's safe to assume that users will not be too confused no matter what you do.

  • 27
    +1 for the linguistic aspect. From your data, I have an impression that big companies what to be projected as not being sexism by placing female first, which is overreacted. If they really aren't sexism, then the options should be written as how they flows from the minds, which obeys linguistic rules.
    – Ooker
    Commented Mar 28, 2017 at 20:43
  • 7
    @TheoreticalPerson When I go to the homepage of Facebook, not logged in, then it definitely is binary for me: facebook.com. Commented Mar 28, 2017 at 21:02
  • 8
    Non-binary gender has nothing to do with western liberal notions about sexuality and gender. The first country to offer a third gender option in the passport was Nepal.
    – gerrit
    Commented Mar 29, 2017 at 9:49
  • 3
    As for your second conclusion (about Silicon Valley preferences), I wonder if that's mostly tied to alphabetical order.
    – user11900
    Commented Mar 29, 2017 at 14:50
  • 10
    Speaking of “the linguistic aspect”, note that in many languages it's nearly impossible to form a grammatically-correct sentence without knowing if the subject is masculine or feminine. This may be the reason the question is asked in the first place.
    – dan04
    Commented Mar 29, 2017 at 23:36

The first thing is to ask if this absolutely necessary within your system?

If you don't need to ask for a gender then don't.

Modern understanding of what gender means is incredibly complicated (there are at least 6 options I can think of) and can also be incredibly personal.

If you really do need to ask (although I can't envision anything outside medicine or insurance that would make it a necessity) the you should simply ask for the users' "Which gender are you?" and try to offer an option that allows your users to with hold that information - something like "I'd prefer not to answer"

  • 76
    +1 for "If you don't need to ask for a gender [or any other thing in your form] then don't." Commented Mar 28, 2017 at 13:09
  • 23
    @Kylos Do you mean biological sex by birth, biological sex by correction or psychological sex? As I said in my answer, there's no need to ask anyone what private parts they have (or want) unless it's for medicine or health insurance Commented Mar 28, 2017 at 13:24
  • 63
    @AndrewMartin "Psychological sex" doesn't exist. Sex is always biological, determined by what you were born as and whether you have a Y chromosome or not. Everything else goes under the gender umbrella.
    – Pyritie
    Commented Mar 28, 2017 at 13:41
  • 33
    @Pyritie there are at several misconceptions in your statement. Pease take it to biology.se for elucidation. You are mixing genital configuration (phenotype) with the genes behind them (genotype) in a discussion that is about gender. In broad XXI century. Commented Mar 28, 2017 at 14:43
  • 31
    It's not that hard to imagine cases where gender is needed. The Stack Overflow Developer Survey comes to mind. As does anyone who wishes to advertise to you. Outside of medicine there is still the fitness industry. I think this answer would be better if it didn't basically deprecate the user for maybe having a reason to need to know.
    – djechlin
    Commented Mar 28, 2017 at 21:58

I recommend matching the order in which your options are presented in the list. You decided to put them in alphabetically (or so it seems), and I see no reason to change the order in the question.

  • 55
    +1 You actually gave an answer to the question instead of going off on a tirade like everyone else.
    – ohmu
    Commented Mar 28, 2017 at 15:39
  • 21
    part of the role of UX is to challenge requirements. not sure where the 'tirades' are.
    – Midas
    Commented Mar 28, 2017 at 15:56
  • 9
    Finally an answer to the question. This kinda seems the best way to do it. I know some other answer advice to use the length of the words but i actually think a user understands easier the alphabetic order. Commented Mar 28, 2017 at 16:02
  • 40
    +1 for the simplicity of this answer. However, most native English people would state this as "male or female" - it's just a consequence of how the phrase sounds linguistically. "Female or male" sounds wrong.
    – adelphus
    Commented Mar 28, 2017 at 17:13
  • 10
    I probably wouldn't make a formal complaint about it either, but I would find it mildly distracting. From a user experience point of view, why stray from convention? If I'm hired for user design, it's not my place to assert my personal ethics on either my client or the users. If they traditionally are trained by a commonly asked question to look for the "male" option first, it is my job to work with their expectations and not against them. I certainly don't want to exploit my work to make a political statement and distract them from continuing the survey or I've failed in my job.
    – iyrin
    Commented Mar 29, 2017 at 10:16

For user-friendliness, the choices should be listed in the same order they are presented in the question. If you ask "Are you male or female?" then A should be male and B should be female. I would present male first; as that is the norm, it won't throw the user off.

  • 22
    +1 for giving the correct order for those two options, and explaining why.
    – Rosie F
    Commented Mar 28, 2017 at 19:24
  • 30
    Precisely. As a male, I am used to the first/left option being more applicable. I expect that if I were female, I would be used to the second/right option being more applicable. If I wanted to make a point about social equality, then I might favor the "progressive" act of placing female first. But if my goal is actual a simple user experience, then commonplace tradition is likely to be the easier experience.
    – TOOGAM
    Commented Mar 29, 2017 at 5:10
  • 2
    I had not considered the form of the question in my answer, but you're absolutely right. Whether you adhere to convention or not, you have to at least be consistent within the scope of your form.
    – iyrin
    Commented Mar 29, 2017 at 9:50
  • 2
    Are you serious about male being the norm? Commented Mar 30, 2017 at 10:23
  • 6
    Verena, it's most common in my experience. It's also so trivial that I doubt either order is going to significantly disrupt the user, but it's uncommon enough that it would stand out if I saw female appear first. I'm born and raised in USA.
    – iyrin
    Commented Mar 30, 2017 at 17:02

Before proceeding further, I would recommend reading the guidance on collecting data on sex and gender (PDF) to ensure you act both legally and ethically. Depending on the laws in your country, you may not be able to collect the data if it's not relevant.

Not only does the guidance suggest people rethink the binary options, but also:

Members must ensure that participants will be able to proceed through any research without being required to indicate male or female options if they do not identify as such.

They provide a checklist, indicating that among other things researchers need to:

Be clear as to which category of information they wish to collect. Sex and gender are related but distinct and researchers must choose the right concept.

Encourage clients to consider whether the demographic information is relevant. This is especially important if considering collecting trans-history but the appropriate level of detail required for a project should always be reviewed.

Having said that:

There are 5 ways to organise things.

  • Location
  • Alphabetical
  • Time
  • Category
  • Hierarchy or Magnitude

Since none of the others could apply with acceptable logic, the only sensible answer is alphabetical and you would of course introduce the options in the same order:

Are you female or male?

  • Female
  • Male

However, that answers the question only as asked, without getting into a debate about the sensitivity of the question's wording.

Now depending on exactly why you're asking, there may be better, less direct options including ones that don't require a specific statement, or 'helpful' options that are not binary.

The remaining questions relate to products usually targeted at women, would you like to continue?

  • Yes
  • No

We want to make sure the following questions are relevant. Please select an option below that describes you best

  • Female
  • Male
  • Prefer not to answer
  • 19
    Personally I don't care who knows my self-identification-gender, but if all you want is a title for mailing to (and a fairly reliable indicator rather than hard personal data), then just ask the customary Mr / Mrs / Ms / Other(inputbox) when obtaining shipping details. Note: this too has its problems if the site has customers world-wide. But at least its far less intrusive and personal for a sensitive few, and honorific-translation errors are more likely to give rise to comedy than tragedy.
    – nigel222
    Commented Mar 28, 2017 at 17:04
  • 3
    @nigel222: But why would you even need a title (unless perhaps it's earned, like 'Dr.', in a professional setting)? Do you think there'd be confusion if your package isn't shipped to Mr/Ms Joe Smith?
    – jamesqf
    Commented Mar 28, 2017 at 17:16
  • 18
    Alphabetical order will become non-alphabetical after localization.
    – Sasha
    Commented Mar 28, 2017 at 17:27
  • 8
    @jamesqf: "why would you even need a title" - Because, for the sake of illustration, "Dear Smith" might not a socially acceptable form of address in a letter, as opposed to "Dear Mr. Smith". Even if you disagree and say that it has become ok in English, or that salutations like "Dear Joe Smith" do not sound "wrong" nowadays, that does not mean the same rules can be applied to other languages. Commented Mar 29, 2017 at 6:35
  • 1
    +1 for the "Prefer not to answer" option. Diversity surveys I have filled in for my employer always include such an alternative, whether it's for gender, sexual orientation, disability status, or otherwise.
    – gerrit
    Commented Mar 29, 2017 at 9:46

Randall Munroe from the xkcd blog has a different approach:

Do you have a Y chromosome?

Don’t Know/Yes/No.
If unsure, select “Yes” if you are physically male and “No” if you are physically female. If you have had SRS, please respond for your sex at birth. This question is relevant to the genetics of colorblindness

The reason is to avoid a question that "doesn’t have a good, clear answer available for transsexuals, intersex people, and people who already know they have chromosomal anomalies". For me, this immediately clears all the debates, avoids asking a sensitive question, but still has the data to explore and analyze. It also provides the reason why the information is important.

  • 3
    +1 because the phrasing describes the reason for the question. But it is still ambiguous in certain cases en.wikipedia.org/wiki/Chimera_(genetics) Commented Mar 28, 2017 at 23:06
  • 3
    It also provides the reason why the information is important. Sure, but that doesn't apply in all cases. "Because we want to sell your data to marketers" is not a reason people are going to like for answering the question of whether they have a Y chromosome. Also, if you just assume having a Y means they are "male", then all kinds of identity verifications might fail if you put in "male" in the backend and they are transgender and have all their other profiles set to "female". Better just ask what their gender is ... and hope the user filled in the same thing everywhere. Commented Mar 29, 2017 at 2:36
  • 4
    This approach specifically worked there because it was looking for a genetic basis for colorblindness. In everyday cases, this is a potentially offensive way to ask: a trans woman is a woman, and her chromosomes are irrelevant.
    – Cascabel
    Commented Mar 29, 2017 at 16:18
  • 4
    This suggestion is unconventional, complicated, and, outside a few specific medical contexts, unlikely to be the most relevant formulation of the question. While at first glance it appears to be a reductionist approach, it fails to recognize that biology is complicated. What's more, it is about as insensitive to concerns of transgender people as the North Caroline state legislature. Commented Mar 29, 2017 at 19:56
  • 4
    This is almost never the right answer unless you're specifically investigating a biological phenomenon, as Munroe was. That's rarely the case for most internet forms. Commented Mar 30, 2017 at 21:47

I think it'd make the most sense to order it by the most likely expected response. Similar to putting "United States" first in a list of countries on an primarily USA-based app, there's nothing wrong with ordering it for ease of use for your particular audience, especially if you target a particular gender demographic.

As a transgender person myself, regarding the nature of the question, it's heartwarming to see these other answers care about being inclusive, and I agree that in nearly every case, limited it to "female or male" would be annoying at best, or offensive at worst. To that, I'd suggest being more specific about what exactly you're going to use this information for:

If it's for a genetic survey, ask for XY, XX, and a field for other to fill in various XXY, XO etc. -- a genetic survey should be interested in intersex/Klinefelter-syndrome info anyway.

If it's for demographic/analytics, push for at least adding prefer not to say so people can opt out for privacy concerns.

If you're using it to make your website more friendly and personalized, then a friendly thing to do would be to ask How should we address you? with Female (she/her) Male (he/him) and preferably a neutral option Neutral (they/them). Yes, it's OK to use singular "they", even recently under the AP Style guide. Facebook does this really well if you're logged in and choose a custom gender (they do a binary choice for new users though):

Facebook Pronoun Picker

Distinguishing between "gender" and "sex" is pointless, because the user will likely answer as their "lived-as gender" regardless of what you ask, unless you craft a long tedious explanation of why biological sex is most important (and even in that case, I'd personally answer with my lived gender anyway because I'll still be skeptical about why you need to know).

  • 3
    +1 for "Which pronoun do you prefer?". It leaves the decision and usage to the end user, instead of saying "what's your gender" and making a decision based on our own lifestyle.
    – cloudworks
    Commented Mar 30, 2017 at 20:28
  • 1
    In theory, the pronoun approach might be problematic under localization (e.g., If the Sun and the Moon had facebook accounts, they would pick different pronoun-genders in Englisch than in German); note that some language might not have a neutral gender or at least not use the neutral plural form in a similar fashion as English does; or the grammatical gender might not as predominantly be related with sex (I don't have a specific langugae in mind here); or maybe some obscure language even has extra pronouns specifically for transgender etc.(again I am unaware of an actually existing such lang) Commented Mar 31, 2017 at 11:43

I assume you want to avoid the word "sex" and limiting gender to only two options may seem offensive to some.

Nevertheless, coming back to your original question, I'd opt for the fist choice, "Are you male or female?". The justification is that in many languages one achieves a sort of pleasant words flow when placing a shorter component before the longer one.

If you have the impression that "male or female" may favour one gender over the other, as an alternative you might thing about that: "ladies and gentlemen" [shorter and longer] sound more natural than "gentlemen and ladies".

  • 45
    It's not that limiting it to 2 genders is offensive to some, it's that it's incorrect. It's like having a 'What is your race: White or Black' field.
    – JonW
    Commented Mar 28, 2017 at 10:29
  • 22
    @JonW Incorrect according to whom? Google has the definition of gender as: "the state of being male or female (typically used with reference to social and cultural differences rather than biological ones)."
    – Henry A.
    Commented Mar 28, 2017 at 14:07
  • 20
    It's incorrect because it isn't accurate for the people using the system. It doesn't matter what Google says is the definition, or what a dictionary says, or anything other than the User. We are UX practitioners, we need to ensure our solutions are fit for use by those users. Limiting it to just Male and Female does not.
    – JonW
    Commented Mar 28, 2017 at 15:07
  • 8
    @JonW You could at least provide a single example (Perhaps you have elsewhere; if so I missed it). If you are thinking of transgenders, that just means that gender identity is not considered; the information collected would be different than that, but not incorrect. For chimeric individuals, there is often a dominant DNA set. The "androgen insensitivity" mentioned elsewhere does not seem to make 2 genders incorrect either. I can think of some ways that someone might need to answer "both", but that is so extremely rare that the form is likely to have worse problems elsewhere first (ex: names).
    – Aaron
    Commented Mar 30, 2017 at 17:21
  • 12
    ... In short, if the question is good for 99.9999% (or more) of humans, it might not be worth it to spend effort on covering the other 0.00001% (or less) who might never even use your form anyway. You are more likely to get a blind (cannot even read your form), far-eastern (your name scheme doesn't work for their culture [no first/last name]), 100 year old (your "birth year" combo list doesn't include 1915) individual which has problems with 3 fields on your form at once. Please identify how the male/female option is actually technically incorrect (not just unpleasant for some). The UX is fine
    – Aaron
    Commented Mar 30, 2017 at 17:29

Take a look at this article https://www.psychologytoday.com/blog/the-science-behind-behavior/201609/how-should-market-researchers-ask-about-gender-in-surveys since it has interesting information about asking gender on a survey, even for cases when the user is transgender.

Here's a paragraph that sums up everything and the best approach (according to the article):

(...) by asking the “What’s your gender?” question and providing only the “male” and “female” options to choose from violates both principles. It is an implicit endorsement of discrimination against transgender and gender non-conforming respondents. And in some cases, the analysis may end up masking differences between respondent groups (because gender was not elicited correctly). Using a five-category question to provide gender that allows respondents to choose from male, female, transgender, and other, categories, and the freedom to write-in a custom response if other is chosen, will increase both accuracy and inclusiveness of market research surveys.

  • 21
    By that standard the question "Are you a student?" should in addition to "Yes" and "No" offer options such as "Admitted but haven't started yet", "Just expelled", "Student at heart" etc. Commented Mar 28, 2017 at 14:55
  • 2
    Article's author seems to be trying to please everyone, but didn't mention any attempts to actually ask the people he's surveying.
    – cHao
    Commented Mar 29, 2017 at 19:15

To answer the direct question, I prefer "male" before "female" as a user.

tl;dr reasons:

  • It's more euphonic.
  • It's conventional.
  • Efficiency is not sacrificed for either user case.

As I am reading through the form, it is much more euphonic to read "male" prior to "female." Go ahead and say them each out loud. "Male-fe-male." Now say "Fe-male-male." The latter is so phonetically ugly that it interrupts my flow as a user. This has been the conventional order of appearance for so long, probably because it is more phonetically pleasing, that to suddenly see it switch would appear out of ordinary and may distract me slightly as a user.

Standards and conventions are great for users because we expect something to be in a certain place and it requires way less effort to find it. Imagine if a new update in your web browser switched the order of the back and forward buttons. There's nothing inherently wrong with it. It's no more difficult to click either one. It might even make more sense for somebody who is used to an RTL language.

But as a western convention, it would be extremely out of the ordinary and not intuitive for your average user. It goes against well established user habits. In your case, the user habit has been established before they were even exposed to your form.

Remember that you want me to fill out your survey. I understand how difficult it can be to find people who are willing to participate. That's why it is important to remove as many barriers as possible that may stand in the way of me efficiently completing your survey. Work with the user and their existing habits. Do not disrupt their flow or break convention unless it is necessary and you feel that it will ultimately serve the purpose of providing a better user experience for the most users and reduce your bounce rate.

As far as efficiency for the user making a selection, I do not believe that selecting the first or second option requires any additional effort over the other. Especially having a conventional expectation of "male" before "female," I can select the option that applies to me after having barely looked at it.

The only circumstance I can imagine that would make selecting one option any less efficient for the user is if they are navigating with the keyboard and selecting the second option requires a single additional keystroke. This is not likely to impede user experience or create any barriers to the user continuing and completing the survey.

As far as user experience goes, you really can't go wrong by sticking to the convention your users are familiar with.

Edit: I focused entirely on the options with consideration for the content and failed to consider the question "Are you male or female" as it introduces the options. The options should definitely be presented in the same order as they are first introduced to the user. It should not matter whether the options are male/female or cats/dogs. A user is learning to interact with your form as they use it. Consistency makes this process much easier and inconsistency leads to confusion and disruption.


Make the page choose a random number in JavaScript. If it's even, display the female choice first; if it's odd, display the male choice first. Also, the query text should be in the same order as the choices. After those two choices, you should have an "other" choice.

This way, nobody can accuse you of bias for putting one choice before the other! Half the time it's one way, half the time the other. As a bonus, the people who want the "male" choice to be first for the sake of tradition are satisfied customers half of the time.

  • 26
    Your testers will hate you for doing this 100% of the time ;) Commented Mar 28, 2017 at 16:43
  • 4
    @DmitryGrigoryev So you make the script respond to #!m or #!f at the end of the URL that will set the order deterministically! This will make the testers happy, and possibly even power users. Commented Mar 28, 2017 at 19:30
  • 3
    @Ooker: Testers tend to hate randomness, cause it makes it a lot harder to do their job and test everything. If you can induce one outcome or the other at will, though, that problem is mostly solved..
    – cHao
    Commented Mar 28, 2017 at 22:18
  • 8
    You can displease all the people all the time.
    – user67695
    Commented Mar 28, 2017 at 22:37
  • 1
    @Ooker: Thing is, no matter when you "implement the randomness", it needs to be tested. And you don't want the developer doing all the testing. (Blind spots, tunnel vision, and confirmation bias are real issues.) So it falls to the testers, and the problem returns. (Aside from that, you have the problem of deciding which order the static content should take. And now we've come full circle.)
    – cHao
    Commented Mar 29, 2017 at 18:59

You should not worry about placing them in any particular order. Just toss a coin and place them randomly (at design time; doing so per each page reload makes little sense). Why?

Because being obsessed so much with the (very immature) idea of "listed first == respected most" just helps the world to continue becoming a worse place to live in and stimulates all kinds of baseless hypersensitivity to imaginary and/or far-fetched insults, like being listed last or first.

You see, in order to satisfy everybody you'd have to reject the very notions of space and time and go quantum, superimposing all the text items denoting different 'gender configurations' (trying your best to guess whatever new term for 17th or 23rd configuration they've just came up with recently) on top of each other in one place, and also making sure that nobody/nomind/nosoul is able to actually collapse this superposition into a state unfavorable for her/him/it/one and thus become insulted; singularity seems to be the goal of all progressive mankind/womankind/humankind/beingkind after all, they gotta love it.

Also, what if someone gets insulted for not being listed first from the bottom or at the center?

P.S. You may also want to inform your users explicitly that the order of items is random, so that they could only blame the Merciless Hand of Providence and not your evil intent.

  • 5
    Your post is rather contentious, but apart from that randomizing the order is probably the worst thing you can do. If for whatever reason someone needs to fill in the form a second time (for example because posting it failed) having a different order from last time without a clear reason, adds a cognitive load. It makes it more likely that the user accidentally gives the wrong answer because, for example, they assume the default option is always the same and skip over the question.
    – nitro2k01
    Commented Mar 29, 2017 at 5:46
  • 3
    Worst thing for bots and inattentive users, yes.. But my point is that is frees you from being accused of lack of political correctness - and if you follow any arbitrary order you're still leaving room for that. You know, there's a war on common sense going on right now.
    – sunny moon
    Commented Mar 29, 2017 at 8:08
  • 1
    @nitro2k01 as he advocates tossing a coin, not using an RNG - I think this is more a way to pick the static order of two items, not to randomly switch up the order on every page load.
    – Baldrickk
    Commented Apr 3, 2017 at 14:07

Just going to leave this wiki article about intersex here.

As such, from a professional point of view, irrespective of whether your question is enquiring about their physiology or orientation, you should leave an option that other than "Male" or "Female".

Typically, this can be as simple as "Other" or "Prefer not to say".

Specific to the 2 images given, I take it your question was meant to emphasize the order to present the options in.

In this case, why not just stick to alphabetical ordering? The "Other" or "Prefer not to say" option should always be last.

  • 3
    Far better to not ask, unless you absolutely CAN'T avoid it.
    – WGroleau
    Commented Mar 28, 2017 at 16:18
  • 2
    Not so sure I understand why you need other if it's for medical reasons. If this question is for medical reasons then knowing if a person has a uterus or a prostate may very well be required information, and other is not an acceptable response.
    – Mayo
    Commented Mar 28, 2017 at 19:44
  • 3
    I would suggest that the default choice should be "Unspecified", listed last. An advantage of the term "Unspecified" is that it will by definition be correct until one of the other selections is specified, and implies nothing about why it was chosen.
    – supercat
    Commented Mar 29, 2017 at 19:51
  • 1
    I like this, but "Other" and "Prefer not to say" are two very different meanings. The first covers e.g. nonbinary people (or people who want to give a more complex answer, which might include, for example, some but not all intersex people and a few trans people who don't feel comfortable giving a binary answer), whereas the second is declining the question. It might be worth including both. But if I was nonbinary, and my only reasonably accurate option was "Prefer not to say," I think I would be a bit annoyed.
    – Bemisawa
    Commented Mar 31, 2017 at 20:04
  • 1
    @Bemisawa - Correct. My intention was AND/OR <non-binary option> not exclusively OR. @ supercat - Unspecified is also fine. One other option is to mark this as an optional question. This would address the use case of asking the question but also would address the scenario where the user does not have to answer the question if he doesn't want to. (This is moot if your use case requires an answer to the question)
    – lohithbb
    Commented Apr 1, 2017 at 10:01

In survey software, options are listed in random order, and the question should not itself suggest any answers in any order.

If you are asking a variety of similar questions, eg:

  1. What is your gender identity?
  2. What is your birth gender?
  3. What is your sexual preference?
  4. For what gender identity are you most often buying our product?
  5. What gender do you feel our product is the best match for?

... etc, then the answers may be randomized per-form, rather than per-question. So half of the people taking the survey will have male-first on all their questions, and half will have female-first, on all questions.

The idea behind this is to prevent any biases that might come from having a fixed order to the questions.

As others have pointed out, though - do make absolutely certain that you are providing options to cover all applicable cases, and do make absolutely sure that you need, and are permitted to gather this information in the manner that you are gathering it.

If displaying their response publicly (if it's a social networking site profile form, for example) then you will need to be even more careful.

Consult your legal team before completing the design phase of this work.

  • 1
    +1 for the legal team. Just did a survey yesterday, including this exact question. The options as presented to me were: 1) Female 2) Male 3) Other 4) Prefer not to say.
    – gerrit
    Commented Mar 31, 2017 at 12:51

Irrespective of what you do, the conversions won't improve even by a single %

Just pick a rationale like alphabetical order and go ahead focus on bigger problems

Not the answer you're looking for? Browse other questions tagged or ask your own question.