Using “HypoTestingTask2Data.sav” dataset on MOLE you are to analyse some results of a small-scale educational research study.
Context of the study
This fictional dataset contains English Literature (english_lit) and English Language (english_lang) scores from tests conducted with pupils in four schools (school), as well as a record of whether or not each pupil is diagnosed with some form of Special Educational Need or Disability (send). The data can be assumed to be a randomly selected sample from the population of each school.
Your task is to conduct an analysis to answer one of the following questions:
- Does School 1 have significantly different results for one of the English tests (language or literature) compared to School 2? You should choose to analyse either the English Literature or English Language results, do not try to report on both.
- Is there any relationship between the school a pupil is attending and whether or not they are diagnosed with some form of SEND?
- Is there evidence to suggest that the four schools in the study have significantly different results on one of the English tests (language or literature)? You should choose to analyse either the English Literature or English Language results, do not try to report on both.
- Is there evidence to suggest that pupils with SEND achieve different results on average to pupils without SEND? Again, you should choose to analyse either the English Literature or English Language results, do not try to report on both.
- Do these pupils, on average, achieve significantly different results on their English Literature and English Language tests?
You must state and justify your choice of test, taking care to explain any assumptions you are making of the data, including the results of any additional tests you need to conduct to check those assumptions.
If you find evidence of a significant result then you should produce either summary statistics (this can include a cross-tabulation, averages etc.) or graphical output to explain your findings.
The main aim of this assignment is to allow you to produce some appropriate statistical output, interpret your results and justify your choices. Ensure that you cite suitable methods texts or similar prior research studies where possible.
- Begin by explaining what question you are wanting to answer from the first page (discuss the context of the study and what the variable is supposed to measure)
- Explain what sort of analysis you will conduct (Will you conduct a t-test? ANOVA? Chi-square test? Why?) Remember to cite some methods literature to justify your choices.
- Explain if any assumptions are required to conduct the test and how you are going to check them.
- Conduct the analysis and put the output (statistics or graphs) into your report.
- Explain (in words) what these results tell you about the schools or the pupils (depending on the analysis)
- Calculate averages, crosstabs, draw graphs or include other output to help you explain your results in greater detail. E.g. which school got different results to the rest, which school seems to have significantly more SEND pupils than you would expect.
- Remember to explain your results in words related to the context of the study, talk about “English literature scores” or “numbers of pupils with SEND” – not just the technical terms like “means” and “variables”.
- You should state the research question you have chosen to answer at the start of the report.
- You should explain any assumptions you make in order to conduct your chosen test, providing the results of any supplementary tests you need to conduct to justify your choice.
- You should explain your findings briefly with the aid of descriptive statistics, tables or graphs.
- You must justify the types of tests, statistics or graphs you use in relation to the context and the type of data which has been recorded.
- You should cite methods texts or other studies that have been conducted in education as justification for your choice of approach.
- There is no strict word limit for this section but I advise that the report should be between one and three pages in length.