Your IA experiment didn’t work? You were hoping to “prove” your theory right and have the same results as the original weren’t you? Never mind, this can teach you a few valuable lessons and it won’t affect your marks. Here’s what you need to do.
You’ll still need to conduct your descriptive and inferential statistics. You’ll also need to explain your results in relation to the background theory/model.
Nothing changes in this section. You still calculate your central tendency (average) and the dispersion (spread). Not sure if you should use the mean, median or mode? This video and/or this post will help.
Remember: you don’t need to justify your choice of statistics. However, you must interpret the results. This means explaining what they suggest about the topic you’re studying.
- Watch: Top 5 Mistakes in the IA Analysis
- Watch: Playlist: All IB Psych IA Videos
- Read: All IB Psych IA Blog Posts
Presumably you had a one-tailed hypothesis, meaning you were predicting one group to score higher than the other. But you got the opposite. You still need to calculate your inferential statistics using an appropriate test (e.g. Mann-Whitney, Chi-squared or Wilcoxon). I recommend the social science statistics calculators.
Here’s what to do if yoru results are:
- Statistically insignificant: Accept your null hypothesis.
- Statistically significant: You still accept your null hypothesis.
You should make it clear that you’re aware your results are the opposite of what was hypothesized. Even if the calculations show significance (p = <0.05), they are “in the wrong direction” and so you accept/retain your null hypothesis.
There’s debate about the “proper” way to handle such findings. However, at this level of studying Psychology it’s enough for students to simply run the inferential test and apply to the hypotheses.
You must make sure the “…findings of (your) investigation are discussed with reference to the background theory or model” to get top marks. Even if your results are in the wrong direction, you need to state this and show the results perhaps contradict the background theory. You can read more about how to write the evaluation in this post.
If you were disappointed your results “didn’t work,” you now have a first-hand taste of “researcher bias.” There’s actually no real reason to be disappointed because you shouldn’t have been trying to prove anything, you were testing a hypothesis. But, it’s inevitable you’re disappointed. You now know how researcher bias can influence psychological studies – psychologists want their experiments to “work” and this can affect their results in many ways. That’s not to say that your results didn’t work because of researcher bias, however. It just shows you know the human emotion that goes into creating studies and how this might affect their design and thus their findings. This could be why Psychology is going through a replication crisis right now, which makes for great fodder in your TOK classes.
Travis Dixon is an IB Psychology teacher, author, workshop leader, examiner and IA moderator.