GoConqr has a new look! Let us know what you think.

Sign up free

Flashcards

research methods booklet 3

megan trent Psychology English 80 cards

research methods booklet 3

Cards in this deck (80)

  1. 1Quantitative data:This is data which is in numerical form. For example, age, how many hours you work in a week, ratings of TV programmes.
  2. 2Qualitative data:This is data in non-numerical, descriptive form. For example this data may include transcripts from interviews, detailed descriptions of what has been observed or the written responses to questions.
    Both forms of data can be used collectively to give deeper insight into research findings. Qualitative data can also be converted into quantitative data through content analysis and then presented in quantitative form such as tables and graphs and analysed by numerical means such as statistical tests.
  3. 3Quantitative vs QualitativeQuantitative vs Qualitative
  4. 4Quantitative vs QualitativeQuantitative vs Qualitative
  5. 5Experiments qualitative or quantitive ?Tend to produce quantitative data as the researchers are very accurately measuring the DV to see the difference between the conditions. However participants may be interviewed afterwards about the study, which will probably provide qualitative data. However these responses may then be analysed further possibly converting them into quantitative responses
  6. 6Observational Methods qualitative or quantitive ?Structured Observations, which usually involve the researchers documenting the observed behaviour in highly structured categories tend to produce quantitative data.
    Less structured observations might just make detailed notes of all relevant behaviours, therefore producing qualitative data.
  7. 7Correlational Analysis qualitative or quantitive ?The co-variables are quantified to plot them onto a scatter graph. Also a Correlation Coefficient is calculated to provide more information about the strength and direction of the correlation. Therefore this method uses quantitative data.
  8. 8Questionnaires qualitative or quantitive ?Closed questions in questionnaires will result in quantitative data, whereas open questions will result in qualitative data.
  9. 9Interviews qualitative or quantitive ?Closed questions in interviews will result in quantitative data, whereas open questions will result in qualitative data. Unstructured interviews are likely to only produce qualitative data as they use very open and broad questions.
  10. 10Case Studies qualitative or quantitive ?Can produce quantitative data, but are much more likely to produce qualitative data as they are using highly detailed and descriptive accounts of the case.
  11. 11Content analysis qualitative or quantitive ?Content Analysis converts qualitative data into quantitative data in order to observe patterns and trends within a data set. This is achieved by categorising and counting the number of times a particular event or behaviour occurs.
  12. 12Primary datarefers to original data which has been collected in order to investigate a specific research aim and has not been published before. It is collected first hand from the participants themselves eg by experiment, questionnaire, interview or observation.
  13. 13Secondary datarefers to data that has already been collected by someone other than the person conducting the research. It has already been subjected to statistical testing and published eg journal articles, books, websites and government statistics.
  14. 14Meta-analysisis a statistical technique for combining the findings of several studies that have involved the same methods of research. This allows for the identification of trends and relationships that wouldn’t be possible with individual smaller studies. The researcher may alternatively carry out just a qualitative analysis where they simply discuss the findings / conclusions of the studies. If using a quantitative approach, this will involve calculating an effect size which is the dependent variable of the meta-analysis, giving an overall statistical measure of difference between variables across a number of studies.
  15. 15Evaluation Primary dataThis is authentic data that is obtained from participants themselves for the purpose of the particular investigation eg questionnaires and interviews can be designed in such a way that they specifically target the information that is required. This means the researcher has control over the data so that it fits the aim and hypothesis of the study.
    It can be a lengthy process and expensive as participants have to be collected and the study has to be conducted and analyzed. Therefore, this means significant effort is required to produce primary data eg conducting an experiment.
  16. 16Evaluation Secondary dataThere may be substantial variation in the quality and accuracy of secondary data, some information may appear valuable initially but may be outdated or incomplete, thus not meet the needs of the study.
    As this data was previously collected by someone else there is uncertainty in the reliability of the data and the way it was conducted. This means the data may have been manipulated unlike primary data.
  17. 17MEASURES OF CENTRAL TENDENCY:Mean, Median and Mode (Measures of central tendency provide information about the most typical value in a set of data, or the value around which all other values are clustered)
  18. 18Mean (The average)All the values of the set of data are added together and divided by the total number of values. Most appropriately used with interval level, continuous data (e.g. time, speed, temperature)
  19. 19Strengths of using the meanUses all scores in a data set - This means that the mean value is representative of all the scores and not just a select few.
    The most accurate measure of central tendency– it uses the interval level of measurement where the units of measurement are of equal size.
  20. 20Weaknesses of using the meanAffected by extreme scores - For example, one extreme score in a data set could vastly affect what the mean score is.
    It may not be a whole number - For example, it is possible the mean score might include a decimal point (e.g. 7.5.)
  21. 21Median(The central value): All the values in a set of data are arranged in order, the middle value is the median. Most appropriately used with ordinal level, subjectively ranked data (e.g. scores in a memory test, ratings of attractiveness)
  22. 22Strengths of using the medianNot affected by an extreme score - This is because the median only takes the middle scores into account when being calculated.
    Easier to calculate than the mean – Unlike the mean, you do not need a calculator to work out the median score
    Can be used with ordinal data - unlike the mean.
  23. 23Weaknesses of using the medianDoes not take all scores into account – Because only the middle numbers are used to calculate the median this means the other scores are ignored making it less sensitive.
    May be unrepresentative if there are few scores – For example, if a data set only includes three or four scores which are very varied, it is unlikely that the median score will be representative of data set.
  24. 24Mode(The most frequent value): The value in a set of data which occurs most frequently. Most appropriately used with nominal level, category data (e.g. numbers of smokers of a certain brand, favourite newspaper)
  25. 25Strengths of using the modeNot affected by an extreme score – When the mode is calculated only the most frequent scores are taken into account, this means extremes scores will not affect the mode.
    Easier to calculate than mean – Similarly to the median you do not need a calculator to work out the median score.
  26. 26Weaknesses of using the modeIt is not useful for small data sets – Similarly to the median if a data set only includes three or four scores which are very varied, it is unlikely that the median score will be representative of data set.
    There may be more than one answer – If more than one score is most frequent it is possible to have more than one mode.
    Less sensitive – it does not use all the scores.
  27. 27Range:Tells us over how many numbers the data is spread. It is the difference between the highest and lowest values in a set of data. To allow for any statistical error you should add 1 to the difference. The range is most appropriately used with the median.
  28. 28Strengths of using the rangeEasy to calculate – No complex calculations are needed to work out the range, compared to the standard deviation.
    Gives a basic indication of the spread of scores – Takes account of extreme values.
  29. 29Weaknesses of using the rangeDistorted by extreme scores – Similarly to the mean, one extreme can vastly affect the range and make the score unrepresentative of the data set.
    Does not indicate how scores are grouped around the mean – Although the range is easy to work out, it does not show much about the dispersion of the score; the standard deviation can tell researcher much more.
  30. 30Standard Deviation:Normal Distribution CurveAn accurate and powerful measure of the spread of scores around the mean value.
    Large standard deviations are obtained when there is a wide spread of scores.
    Small standard deviations are obtained when the scores are closely clustered around the mean.
    Essentially the larger the SD the greater the spread of scores (greater variance); smaller the SD the smaller the spread of scores (less variance), meaning the Mean is a more representative MCT for that data set.
  31. 31Strengths of using the standard deviationMost sensitive measure of dispersion – as all scores are taken into account.
    It allows for the interpretation of individual scores – you can see where a score falls in relation to all others eg anybody with an IQ of 121 is in the top 5% of the population between +2 and +3 sd’s
  32. 32Weaknesses of using the standard deviationCan be difficult to calculate – The standard deviation requires a fairly complex calculation, which is time consuming compared to the calculation of the range.
    Relies on a normal distribution – can be less meaningful if data is not normally distributed.
  33. 33Calculating Standard Deviation1. Add all the scores together and divide by the number of scores to calculate the mean.
    2. Subtract the mean from each individual score.
    3. Square each of these scores.
    4. Add all the squared scores together.
    5. Divide the sum of the squares by the number of scores minus 1. This is the variance.
    6. Use a calculator to work out the square root of the variance. This is standard deviation.
  34. 34When to use each Measure of central tendency and measure of dispersionWhen to use each Measure of central tendency and measure of dispersion
  35. 35Distributions: Normal and skewedWhen certain characteristics are measured such as height, weight or IQ of large numbers of people within a population, the frequency of the measurements should form what is known as a bell shaped curve. This is known as normal distribution and is a symmetrical shape when plotted on a frequency polygon (line graph). Normal distribution is found when the data is continuous, each score is the result of randomly distributed effects and there are large numbers of scores drawn randomly from the population.
    There are several ways the data can be checked to see if it is normally distributed: 1. Examine visually to see if most scores are clustered around the mean. 2. Calculate measures of central tendency to see if the mean, median and mode are the same. 3. Plot the frequency distribution on a histogram to see if it forms a bell shaped curve.
  36. 36Normal Distribution (bell shaped curve)Normal Distribution (bell shaped curve)
  37. 37Skewed DistributionSkewed Distribution
  38. 38Skewed DistributionSometimes data is not symmetrical and can become skewed by outlying scores. A positive skew is where most of the distribution is concentrated towards the left of the graph therefore most participants scored on the low side of average. A negative skew is the opposite. When data is skewed the mode remains towards the highest peak, the median comes next and the mean has been dragged the furthest from the centre.
  39. 39Skewed distributioncan result from:
    1. A small number of scores being taken
    2. A biased sample 3. One end of the scale has a cut off point eg reaction time as at the fast end it is not possible to better 0 seconds.
  40. 40PercentagesThese are a type of descriptive statistic that shows the rate, number or amount of something within every 100. Data can be converted into percentages by multiplying them as a factor of 100 eg:Percentages
  41. 41Converting a percentage to a decimalTo convert a percentage to a decimal, the % sign is removed and the decimal point is moved two places to the left eg 37% is 37.0 then becomes 0.37
  42. 42FractionsFractions are part of a whole number such as ½ or ¾. If results need to be presented as a fraction the following calculations will be done. For example, to work out what fraction of participants are in condition A if there are 40 in condition A from 120 participants in the whole study. 40 is divided by 120 = 40/120, then the fraction is reduced by dividing the top number and the bottom number by the lowest number that divides evenly into both 9the lowest common denominator). This result in 40 being a fraction of 1/3.
  43. 43Converting a decimal to a fractionIf there are two decimal places then you divide by 100 eg 0.24 becomes 24/100. If there are three decimal places then you divide by 1000 eg 0.875 becomes 875/1000. The fraction can be further reduced by finding the lowest common denominator, the biggest number that divides evenly into both parts of the fraction. In the second example, both numbers divide by 25 so that the fraction then becomes 35/40.
  44. 44Using ratiosInformation can also be expressed as a ratio. This could be a “part to whole” ratio eg if 13 out of 20 participants obeyed in an experiment, it would be 13:20. It can also be written as a “part to part” ratio ie if 13 obeyed and 7 did not obey, it would be expressed as 13:7. The lowest common denominator should always be found when expressing the ratio.
  45. 45EstimationsIt may be necessary to comment on the central tendency or dispersion of a set of data which may require estimating the answer. If, for example, the highest score in a set of data is 206 and the lowest 59, an estimate of the range using rounded figures of 200 and 60 would be 140. Estimations of this type are called order of magnitude calculations.
  46. 46Use of significant figuresA long number is rounded off to the nearest thousand or million eg 432,765 to two significant figures is 430,000. When many numbers occur after a decimal point we may round this off to 1, 2 or 3 significant figures eg 0.003245 to two significant figures is 0.0032. When the final digit is 5 or above, the previous digit is rounded up and rounded down if the final digit is less than 5.
  47. 47Order of magnitudeWhen dealing with larger numbers if might be clearer to give two significant figure and then say how many zeros there are, this is known as the order of magnitude. For example, 8,600,000,000 is would be presented as 8.6 x109 (9 represents how many places we have moved the decimal point)
  48. 48Summarising data in a tableOne way of representing data is in the form of a summary table. When tables appear in the Results section of a scientific report, they are not raw scores but have been converted into descriptive statistics
    It is standard practice to include a summary paragraph beneath the table explaining the results. Eg: It appears that on average people were able to recall more words when in a quiet room and that noise negatively affected memory. Scores were however more dispersed in the noise condition showing some people were much more affected by the noise than others whereas in the quiet condition, scores were much more clustered and similar.
  49. 49Bar ChartsBar charts are used to represent the frequencies of non-continuous (or category) data. The x axis illustrates the groups or categories and the y axis illustrates the frequencies. There are gaps between the bars to illustrate that the data is not continuous.
  50. 50HistogramsIn a histogram the area of the bar (not the height), shows the frequency of the data. Histograms are typically used when the data is in groups of unequal width. In a histogram the bars touch each other as they are used to represent continuous data rather than discrete as in a bar chart. The x axis is made up of equal sized intervals of a single category eg scores in a Maths test. The y axis represents the class width of each interval (which are likely to vary in size). If there was a zero frequency for one of the intervals the interval remains but without a bar.
  51. 51Line graphs (Frequency polygon)These also represent continuous data. The graph is produced by drawing a line from the mid-point top of each bar in a histogram. With a line graph, two or more frequency distributions can be compared on the same graph.
    Example graph. Comparison of time take to track a large or small target on a computer screen
  52. 52ScattergramsScattergraphs show the relationship between two COVARIABLES. They are used to represent the CORRELATION. between two sets of continuous data. See Correlations (in Booklet 1) for more information.
  53. 53Statistical TestingThe purpose of statistical testing is to see whether differences or relationships have occurred due to chance or whether the IV has had a significant effect on the DV (in an experiment).
  54. 54Probabilityis identified by the letter p and is a measure of how likely the difference or relationship between two sets of data is a real (significant) difference/relationship or whether it has just occurred by chance. Probability is expressed either as a percentage or on a scale from 0 to 1. The accepted level of probability in Psychology where a result will be regarded as significant is 5% or 0.05. This means that there is a 5% probability that the results in a particular study occurred by chance and a 95% probability that the results were due to the IV (if it was an experiment).
  55. 55Significance levels:Significance refers to whether the results of the study will be accepted as down to chance or as an effect of the IV. When looking at the size of the difference between conditions, there has to be an agreed point at which it will be accepted as significant. There is never a 100% certainty that differences/relationships are beyond the boundaries of chance so it is impossible to prove something beyond all doubt. The accepted cut off point in Psychology to regard findings as significant is 5% which means you are 95% certain your results were not down to chance, leaving a 5% possibility of error. 5% or 0.05 is therefore known as the accepted level of significance. In some circumstances researchers may need to be even more confident that their results were not due to chance eg where there could be a serious negative effect on individuals such as drug testing. They therefore may use a 1% or 0.01 level of significance as this leaves only a 1% chance of error.
  56. 56There are a large number of statistical tests all of which have their own specific criteria for use. One test is the SIGN TEST.This requires a repeated measures design, at least nominal level data and is used when looking for a difference. To calculate a sign test, follow the instructions on pg 300 of Lawton et al or pg 199 of Flanaghan et al.
  57. 57Once the observed / calculated value has been obtained, this is compared tothe critical table value for the appropriate statistical test. When looking up this value the researcher needs to know whether a directional or non-directional hypothesis was chosen and the number of participants in the study.
  58. 58For some tests to be significant, the observed value has to begreater than the critical value but for the Sign Test it has to be equal to or less than the critical table value.
  59. 59Interpretation of significance:Once you have calculated your observed value you need to compare it against a critical value (see the appropriate table below). If your observed value is equal to or less than the critical value at the 5% level of significance, you can accept your experimental/alternative hypothesis and reject the null hypothesis. If your observed value is greater than the critical value at the 5% level, you must accept the null hypothesis and reject the experimental/alternative hypothesis.
  60. 60Nominal data –FREQUENCIES IN CATEGORIES. This involves placing participants into categories and counting the numbers within each category. It is the crudest, most uninformative type of data. Eg how many people said yes v no
  61. 61Ordinal data– RANKED SUBJECTIVE SCORES. This is ranked data eg who finished 1st, 2nd, 3rd in a race etc. Rating scales are also regarded as ordinal data. It gives more information than nominal data but is not as informative as interval as we do not know the precise distance between each ranked place.
  62. 62Interval / ratio data –FIXED/AGREED UNITS WITH EQUAL INTERVALS. This involves standardised measurements on a scale of equal intervals eg time, weight, temperature. It is the most accurate and informative measurement. Interval and ratio are classed together but the difference is that ratio has an absolute zero point (eg weight).
  63. 63SELECTING YOUR STATISTICAL TESTSELECTING YOUR STATISTICAL TEST
  64. 64SELECTING YOUR STATISTICAL TEST what to ask yourself?1. Are you looking for a difference or a correlation?
    2. What is your level of measurement?
    3. What is your experimental design? (Experimental design is not relevant to a correlational study but is the third criteria necessary when choosing stats tests for experiments).
  65. 65Once you have calculated your statistical test, you then determine whether your results are significant by comparing your calculated value to the appropriate critical table value. You then givea statement of significance either accepting your experimental/alternative hypothesis or rejecting the null hypothesis or vice versa.
  66. 66YOU HAVE TO KNOW HOW TO CALCULATE THESIGN TEST.
  67. 67SIGN TEST:Looking for a difference; Related Data (requires a repeated measures (or matched pairs) design), at least nominal level data.
  68. 68EXAMPLE SIGN TEST CALCULATION:This set of data represents test scores at the end of Spring and the beginning of the Autumn terms. The hypothesis is that the summer break means a significant drop in test scores.
    Experimental Hypothesis: Test scores in the Autumn term will be significantly lower than test scores in the Spring term.
    Null: There will be no significant difference in test scores between the Autumn term and Spring Term.
  69. 69EXAMPLE SIGN TEST CALCULATION: STEP 1Subtract the 2nd set of data (Test 2 Autumn) from the first set (Test 1 Spring). If the result is a positive number, put a + in the final column. If the result is a negative number, put a – in the final column. If they are the same, then put 0 and ignore this participant in calculations from now on.
  70. 70EXAMPLE SIGN TEST CALCULATION: STEP 2 –Calculate the sum of the SIGNs (basically count the number of positives and negatives!).
    ignore any 0 values
  71. 71EXAMPLE SIGN TEST CALCULATION: STEP 3– The calculated value (S) of the sign test is smaller of either the positive or negative signs from Step 2.
  72. 72EXAMPLE SIGN TEST CALCULATION: STEP 4To be able to determine whether the results are statistically significant, you need to compare the calculated value to the critical value. Each statistical test has its own set of critical value tables. On the following pages is the critical value table for the Sign Test.
  73. 73In order to look up the appropriate critical value for this data you need to know:1. Whether the experimental/alternative hypothesis was one tailed (directional) or two tailed (non-directional)
    2. The level of significance (this will nearly always be p≤0.05)
    3. n = The number of participants/results. This is the total number of participants minus any whose difference was 0 (rather than +/-).
  74. 74EXAMPLE SIGN TEST CALCULATION: STEP 5– Compare the observed/calculated value of S against the Critical Value. In this case S must equal or exceed the critical value
  75. 75EXAMPLE SIGN TEST CALCULATION: STEP 6If the results are significant the experimental/alternative hypothesis is accepted and the null hypothesis is rejected. If the results are not significant the experimental/alternative hypothesis is rejected and the null hypothesis is accepted.
  76. 76Type 1 Erroris when a result is reported as significant when in fact it occurred by chance. This means a FALSE POSITIVE is reported. This would mean the experimental/alternative hypothesis is accepted and the Null hypothesis is rejected in error.
    It is more likely to happen when using a less stringent Level of Significance (E.g. 0.1), as there is more room for the results to occur by chance.
  77. 77A Type 2 Erroris when a result is reported as NOT significant when in fact it was. This means a FALSE NEGATIVE is reported. This would mean the experimental/alternative hypothesis is rejected and the Null hypothesis is supported in error.
    It is more likely to happen when using a more stringent Level of Significance (E.g. 0.01), as there a greater likelihood that real effects are excluded as occurring by chance, when in fact they were showing a real effect.
  78. 78Using the 0.05 level of significance creates a balance betweencommitting a Type 1 and a Type 2 error.
    This is why it is selected as the conventional level of significance in Psychology.
  79. 79Meta-analysis AdvantagesMeta analysis is an analytical technique designed to summarise the results of multiple studies
    By combining studies, a meta-analysis increases the sample size.
  80. 80Meta-analysis DisadvantagesHigh chance of researcher bias

Information

Description
Psychology (Paper 2) Flashcards on research methods booklet 3, created by megan trent on 30/01/2019.
Created by
megan trent
Views
0
Rating

Feedback

What is it about?

What do you think of the new look?

How do you like it?