-2

I just finished transcribing my interviews for my PhD thesis and I got a total of 200 pages as data to analyse, is this enough? I only transcribed what is relevant to my research scope. Also, English is my second language. One of my friends told me that it is not enough,

closed as off-topic by Roland, Anyon, Richard Erickson, Tommi Brander, Dmitry Savostyanov Jul 23 at 15:42

This question appears to be off-topic. The users who voted to close gave this specific reason:

If this question can be reworded to fit the rules in the help center, please edit the question.

  • 8
    This should have been decided by conducting a power analysis prior to data collection. – Roland Jul 23 at 12:29
  • I'm in the social science faculty – Mahran Jul 23 at 12:32
  • 6
    That's not a valid excuse. – Roland Jul 23 at 12:36
5

Hopefully you have an advisor. That is the person (and probably the only person) who can answer such a question. Fifty interviews may be plenty or it may be woefully inadequate. Four pages per interview may be plenty or it may not be nearly enough.

Ask someone who can actually see your data and who has the experience to make such a judgement in your field.

However, you can at least do preliminary analysis. That might guide you to understand if you need more - or different - data to answer the research question(s).

  • Thank you for your reply, I'm interested to know other experiences in this regard and how much they had had after finishing their transcriptions. – Mahran Jul 23 at 12:33
  • 3
    This varies too much be both field and by the nature of the questions you are researching. No numeric answer from strangers could actually provide you any guidance. – Buffy Jul 23 at 12:34
2

Note: My background is Statistics.

Whether you have enough data to analyze, in other words, have a large enough sample size, depends entirely on what you want to do with your data, and what your data is.

One of the things you can do to find out if you have enough is a power analysis. Power is the probability that you will reject the null hypothesis (i.e, the status quo) when the alternative hypothesis (the thing you are most likely trying to show) is true. In simple terms, power is the probability that if your hypothesis is true, then you will be able to give evidence in support of it given your data.

Since I don't know your data's distribution, I can't really give an idea on how to calculate the power for your specific case, but if you have access to SAS software, or R Studio (free to download) you can use either of these to do the calculation for you.

Generally, this is done before the study even begins, but it can be done after the study is concluded. This is called a post hoc analysis. Generally, it's not a good idea to alter your study based on a post-hoc, since it makes your sample no longer random.* However, if you fail to mathematically support your hypothesis, but your power was only 0.1 to start with, you can simply say your sample size was not large enough to detect the effect you were looking for, and that for future research you plan to increase your sample size.

Before I go, I feel it's worth it to add that your sample size as it currently stands appears to be 50, and not 200. It seems to me that each interviewee is your experimental unit, and even if each page was taken at a different time, these would count either as duplicates or replicates.

*The reason that altering your study makes your sample no longer random. You start with a population of size N, from which you take a sample of size x. The probability of being selected is x/N. You then decide you need more data, so you exclude the original x, so now your population size is N-x. (This assumes the population does not change size). You need y more for your sample, so now the probability of selection is y/N-x, which can never realistically be equal to the first probability of selection, x/N. (I have the proof for the last claim, but I feel it is too far outside the scope of this question to include here. If you are interested, feel free to ask for it in a comment.)

  • 1
    It is also possible, and not stated, that the analysis required isn't statistical. And even possible that the sample taken isn't random. – Buffy Jul 23 at 14:33
  • 1
    @Buffy. True, but inherently the question of "is X enough for [fill in blank]" is statistical. I gave this answer because I personally know people in Social Sciences who could use it if they had a similar question. – GridAlien Jul 23 at 14:44
  • I wasn't complaining, only pointing it out. – Buffy Jul 23 at 14:48
  • @Buffy, my bad :P. – GridAlien Jul 23 at 14:49
  • Well, no, so you can reel in your tongue. ;-) – Buffy Jul 23 at 14:50

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