How experiments can be used to collect information in social research. Learn how surveys, such as interviews and questionnaires, can be used to collect data in social research. Study how content analytics is used to collect data in social research.
It is the linking of the results of the data analysis with the research hypothesis, with the theories and with already existing and accepted knowledge.
Types problems that we could have with the interpretations of certain specific data: Attenuation of the measurement scale. As are to be interpreted performances that systematically reach or can never reach, the limits of the scale of measurement. This problem can be solved by doing a pilot study, detecting these flaws and scaling up the new interpretation.
Ceiling effect. If we always touch the highest scores. Floor effect. If we always touch the lowest scores. Regression to measure. It is an unwanted phenomenon that appears in almost all investigations when a quantitative judgment is requested. It is the tendency to emit responses close to the mean or central values when high extreme evaluations are requested. It can lead us to wrong conclusions.
The results must to be interpreted regarding: The magnitude of the effect obtained and the trends or regularities observed. Compare these results with those obtained by other researchers in similar studies. Clear conclusions of the work carried out.
Data collection: Through systematic observation, surveys and experiments. In natural environments (field study) or in artificial environments (Situations created by the researcher). Data analysis Factors to take into account when performing the four data analysis tasks: You have to decide, although we suggest the double environment: Descriptive Statistics. If we stay in the sample. Inferential Statistics. If we want to infer towards the population using probability. Variables measurement level: Interval or ratio measurement level. Try to measure at the highest possible level, as these include the low ones, but not the other way around. Problem that has arisen and the way in which the data has been collected. A balance must always be made between what is possible and what is convenient, so as not to be inundated with different analyzes. It is advisable to carry out a systematic "analytical" pluralism: Systematicity implies that there must be a detailed plan with specific objectives for both collecting and analyzing data.
Pluralism (any form of research has its limitations. These can be minimized by optimizing the analyzes, for which it is necessary to seek multiple and plural forms of analysis. This plurality includes those referring to non-empirical data and purely mathematical or theoretical developments. Chores of data analysis: Ways to summarize the data. Have indices that summarize different aspects of the distribution. Central tendency indices. They indicate the center of a distribution.
Calculate:
- The arithmetic mean: We add the scores and divide them by the number of them. Ex. (31 + 31 + 25 + 28 + 30) / 5 = 29 The mode: The most frequent observation is 31
- The median: Sorting the scores, the middle score is 30. Indices of variability or dispersion. They indicate how scattered the data of the variable are.
- Variance or biased variance. Calculating the differential scores (subtracting the mean of each score), squaring them, adding them and dividing them by the number of them. Ex. S2s = / 5 = 5.2
- Unbiased variance. We divide the number of cases minus one: Ex. VI = / (5-1) = 6.5
- Unbiased standard deviation. Taking the square root of the unbiased variance (VI) Ex. DTI = Ö VI = Ö 6.5 = 2.55
- Skewed standard deviation. Taking the square root of the Variance or biased variance (S2s) Eg Ss = Ö S2s = Ö 5.2 = 2.28 Total width of the distribution. If the minimum value is subtracted from the maximum value eg AT = 31 - 25 = 6
- Asymmetry indices. Is it a symmetric distribution of scores? Subtracting the mode from the mean and dividing this difference by the skewed standard deviation. As = (29 - 31) / 2.28 = -0.88 If it is less than zero, that is, negative (there are more high scores than low) If it is greater than zero, that is, positive (there are more low scores than high)
If it is zero, it is symmetric (one part of the distribution reflects the other). Aiming indices. Is it a flattened score distribution? Looking for patterns (regularities or differences) in the data. One of the best ways is the graphic representation. Forecasting results based on data. Predictions exploiting your relationships. When a pattern is recognized the best way to summarize it is by means of a function. Although it does not go through all the points, it offers us a simpler, if incomplete, way of describing the data as well as the nature and intensity of the relationships between them.
Generalizing to the population from the sample. Generalize the above results to fields broader than those of the initial sample from which We start by making inferences to the population with the help of descriptive data analysis applying the probability. We go through inferences to generalize towards population results.
This article is merely informative, in Psychology-Online we do not have the power to make a diagnosis or recommend a treatment. We invite you to go to a psychologist to treat your particular case.
