Download presentation
Presentation is loading. Please wait.
1
Hypotheses Hypothesis Testing
2
Where do hypotheses come from..?
Identify the problem Formulate (precise and clear) research question Generate hypothesis (or problem statement) to test TEST using data
3
Sample or population? A sample is a proportion of some larger population The way the sample is selected really matters: random (i.e. all units in the sample have an equal chance of being selected) or non-random (some units in the sample have an increased chance of being selected compared to others) Sampling error is a measure of how well the sample represents the whole population from which it is drawn… the higher the sampling error the less like the overall population the sample is e.g. if I took a sample from this room and only selected women or people wearing black the sample would be less representative than if I randomly selected it giving everyone an equal chance of being included in the sample Generalizability: when the sample matches the population as closely as possible there is a higher degree of generalizability so you can (tentatively) apply your findings for the sample to the whole population…
4
Hypothesis… ‘an educated guess’
A tentative answer to a research question; a statement of (or conjecture about) the relationship between variables being studied Hypotheses should be testable statements about relationships often a prediction that, if confirmed will support a theory Null hypothesis: The hypothesis that is directly tested in hypothesis or significance testing. You want to be able to reject the null hypothesis – typically the null hypothesis is that two or more variables are not related or two parameters are not the same… e.g. Data specialists in policing role are equally likely to be women or men Alternative hypothesis: any hypothesis that does not conform to the one being tested – the logical opposite of the null hypothesis… e.g. Data specialists in policing are not equally likely to be women or men
5
Hypothesis testing Classic approach is to assess the statistical significance… it involves comparing empirically observed findings from a sample with what you would theoretically expect to find if the null hypothesis were true. To do this, the probability of the observed outcome, if the observed outcome were only due to chance is calculated… e.g. observe the number of women and men in this sample of Data Specialists in Policing and compare to our null hypothesis that there is no difference (i.e. you would theoretically expect 50% to be women and 50% to be men)… Using data from registrations: 24 women; 11 men in our sample Reject the null hypothesis
6
Criteria for a good hypothesis…
Declarative form (not a question) Posit an expected (theorised) relationship between variables e.g. Data Specialists and gender; crime and poverty; violence and prior criminal record, etc Based on theory, literature, prior research, valid and reliable knowledge Brief and to the point; direct and explicit (not a fishing trip!) Testable – have data (appropriate quantity and quality) which truly reflects the variables in the hypothesis
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.