Research Questions and Answers
These are Research Questions and Answers discussed by members of the Datarmatics groups.
Godwill Wilcox: Good morning all. How can we validate the reliability of a questionnaire? Students often write, validation of instrument and reliability of instrument. Is it correct?
Favorite Ilo: When an instrument is developed by a researcher or the researcher adopted an instruments, the researcher takes the instrument to a experts in this field for validation. It the experts that will establish the validation by rephrasing some of the item if the need arises.
Godwin Bupo: Good morning all. I have a question. Can Chi-Square be computed for data on interval scale? I have always thought that Chi-square is used when the data is on a norminal or ordinal scale. Please can someone explain to me.
Ellatechy: You can convert data on interval scale to nominal or ordinal scale before applying Chi square.
Good evening my noble colleagues. Can someone help me with the form of how a research development is written? – ABEA MENEGBO
QUESTION: How will you know data cleaning has not affected your objectives?
QUESTION – PATRICK: Ok but does data cleaning address systematic or random errors or both or none ? 1.What will you do if a set of instrument sent to you for analysis has a lot of missing entries? 2. Is there any instruction you can issue to the software like SPSS, Eviews, etc to solve the problem? 3. What will you do if the missing data are many? Also additional question how does it apply in blinded studies. NDUKA WONU
ANSWER – NDUKA WONU
If you are using datasets with categorical variables you need to clean them by getting rid of the non-response categories like ‘do not know’, ‘no answer’, ‘no applicable’, ‘not sure’, ‘refused’, etc. Usually non-response categories have higher values like 99, 999, 9999, etc (or in some cases negative values). Leaving these will bias, for example, the mean age or your regression results as outliers. In the example below the non-response is coded as 999 and if we leave this the mean age would be 80 years, removing the 999 and setting it to missing, the average age goes down to 54 years. Age Frequency 88 2 90 3 92 4 93 1 95 1 999 38 Total 1373
Outliers affect results by inflating the estimates
ANSWER – KELECHI KENNADY
With outliers in your dataset, the assumptions of normal distribution, homoscedasticity and serial correlation will be violated, so you cannot reliably apply most statistical techniques This will significantly affect your result and your research objectives may not be achieved One way to detect outlier and other errors in dataset is through descriptive statistics
QUESTION – PATRICK
If a data is skewed does it mean there an outlier ? So differentiate
ANSWER – KELECHI KENNADY
When you see that the kurtosis coefficient of your data is too large, then you can suspect that there is outlier in your dataset Skewness may indicate leverage, which is also a type error in dataset However, not all outliers or leverage is bad There are good and bad outliers Check your dataset to trace the source of the outlier In other wards this should have been addressed in exclusion criteria ? – PATRICK
ANSWER – KELECHI KENNADY
We don’t avoid bias but removing bad outliers can minimise bias What about exclusion Criteria? PATRICK
ANSWER – KELECHI KENNADY
If the outlier event didn’t occur, then there is good reason to believe that the outlier is bad But if the outlier event actually occur, simply explain the event in your analysis The second part of this seminar will focus on dealing with data errors, outliers and missing observations in SPSS and EViews QUESTION – PROF. OGULU
Thanks, Ken Nnaji. Please recommend at least one very good textbook which treats the subject of data cleaning. It is extremely important in data science/ Big Data Analytics (BDA).
ANSWER – KELECHI KENNADY
“Data Cleaning by Maletic and Marcus(2000)”
REMARKS – raffyb
I am particularly thrilled by the concept of Data cleaning.
Really learning
+234 806 669 7679: Good evening all.pls l need example of a topic that could use design in qualitative research and five research questions to guide the study
Topic: Impact Assessment of Universal Basic Education (UBE) Teacher Training Programme on Student Learning Outcomes in Basic Science.
Draft Research Questions(RQs):
RQ1: What is the opinion of school administrators on the impact of UBE teacher training programme on the learning achievement of students in basic science?
RQ2: How might we describe the impact of UBE teacher training programme on the attitude of students toward basic science?
RQ3. What is the opinion of UBE trained teachers on the impact of the intervention programme on their teaching approaches?
RQ4. What is the assessment of non-UBE trained teachers on the impact of the intervention programme on student attitude toward basic science?
RQ5: How might we describe the opinion of non-UBE trained teachers on the impact of the intervention on student learning outcomes?
RQ6: To what extent has the UBE intervention impacted on the learning progress of students as perceived by learners whose teachers benefited from the programme?
RQ7: How might we describe the impact of UBE intervention project on student learning progress as perceived by learners whose teachers did not participate in the training?
Is it possible to correlate teache motivation(100 teachers were sampled) and student’s achievement(200 students were samples) – AMAKA
NDUKA WONU: To do that first obtain teh motivation score of each teacher using the instrument. Also obtain the achievement scores of every student in teh classes taught by each teacher(eg Teacher named John). Then calculate the average(mean) of all the achievement scored of students taught by Mr John. This single mean score should be matched with the motivation score of John obtained from the motivation questionnaire. You can now correlate the achievement scores(Y) with the motivation scores(X). You should have a total of 100 rows with pairs of scores
So long as the students are taught by the teachers as described above, the students must outnumber the teachers.
Note though that there is no such thing as correlation sampling procedure.
Can I have a clear meaning or explanation of One-group pretest post test design – AWUSE PROMISE
AWUSE PROMISE: One group pretest-post-test means taht you want to work with only one group. You first administer a test(pretest) to the group, get the score of thier performance. Then give them treatment(the independedn variabel being manipulated) then give them another test(post-test). Check out for difference in performance. If the deffrence is as a result of the treatment.