Sampling Flashcards
Define sample surveys
Data are collected from a fraction of the population in order to make inferences about the total population from which the sample was take
Name four importance of samples
It is fast For research Cost effective Due to limited resources(manpower) Saves time
Name five characteristics of a good sample
- (i) Representative,
- (ii) Selected at random to minimize bias,
- (iii) Large enough to increase precision
- (iv) Covered adequately and
- (v) used with a good research design.
Name the types of sampling with their subsets
Probability sampling. –Simple Random Sampling (SRS) –Systematic sampling –Cluster sampling –Multistage sampling
•Non Probability sampling
–Quota sampling
–Snowball sampling
Difference between parameter and statistics
Standard deviation on a sample- statistic or the descriptive numerical measures computed from a sample selected from the population
Standard deviation of a population- parameter Or the descriptive numerical
Measures computed from the population
Define accuracy and precision
Accuracy- how close a measurement or an attempt is to the actual value
Precision- how consistent results are regardless of proximity to the actual value
Something can be precise but won’t be accurate
Define sampling
Process of selecting a number of study units from a. Defined study population
Define probability smoking
I
nvolves random selection procedures to ensure that each unit of the sample is choosing on the basis of chance
Define sample frame
The entire list of all the sampling members or units in the population from which the sample is to be taken.
Define sampling unit
The individual members of the population ie the units which constitutes the sampling frame
Define simple random sampling
The method ensures that each member of the population has equal chance of being selected into the sample.
- There are two ways of obtaining a SRS
- (i) BALLOTING and
- (ii). use of RANDOM DIGIT TABLE
Define systematic sampling
It selects the sample at intervals,
- using sampling interval ‘k’ and selecting every kth item in the list
- starting with an item between 1 and ‘k’ selected at random.
- The method is useful when there is a list of items in the population or if
- elements or items are arranged in rows like houses along a street. or
- patients arriving at a health service
Define cluster sampling
The sample is selected in groups instead of individual sampling units.
- Heads of households in villages.
- The villages are selected at random and within each village, all the sampling units are used.
Stratified sampling definition
STRATIFICATION
•Stratified sampling is used to separate heterogeneous population into homogeneous strata for sampling. within each stratum, any of the above sampling procedure can be used.
–The sub sample size can be selected according to the proportion i.e. proportional allocation, or equal sample size, or the use of optimum allocation.
Define multi stage sampling
•The final sample is obtained by random sampling of bigger units and then further sub sampling of smaller units within the selected bigger units of the previous sampling –For example, •Regions, •Districts, and •Villages.
Define quota sampling
The design divides the population into groups, the quota of subjects for the study are calculated to make sure that the proportion in the sample should agree with the corresponding proportion in the population.
State three disadvantages of quota sampling
The choice for the sample members are left to the interviewer
•Problem of selection bias
•No means of calculating the sampling error
•Conceals the problem of non response
Define snowball sampling
Snowball technique
–Relies on previously identified group members to identify other members of the population. As names are added, the sample grows like a snowball:
–A study of illicit drug users, illegal aliens, prostitutes and community opinion leaders.
Define purposive or judgemental sampling
This is a deliberate non-random method of sampling, which aims to sample a group of people or setting with a particular characters
–Drivers at a lorry station.
Name three ways a questionnaire can be designed for
it can be designed for mail,
•telephone or
•person to person interview
Name one demerit and merit of telephone interviews
Need good telephone net work and use very few questions.
Name one merit of person to person interview
Most frequently used method.
•Need good questionnaire not too bulky.
• The advantage of person to person interview is being there to explain points.
Name one demerit and two merits of the mail questionnaire
has the advantage of over coming sensitivities to private inquire.
–It is necessary; to have clear, simple questions.
–Good postal system and
–incentives to mail back the responses i.e. by including stamped envelopes.
–It has a disadvantage of poor response rate.
Under Qualitative research
Name three points in the Interview guide for
•Focus Group Interview
•Key informant interview
Take notes
•Record the sessions
•Can use text-based Beta for analysis of themes
Name five guidelines to making a questionnaire design
- Care must be taken to limit the number of questions to the information, which is being sought.
- It is important to spell out precisely what is to learnt from each question.
- Avoid bias in the way the question are framed. Sometimes instead of designing one’s own questions on every topic,
- one can use standardized questions that have seen developed over the years, this also allows for comparability of surveys.
Name two advantages and disadvantages of closed ended and open ended questions
•The questions can be open ended or close ended.
•Open ended questions promote confidentiality, respondents are not forced to answer Yes/No to any embarrassing questions spontaneous,
•freedom and variety of responses can be achieved.
•However, it is difficult to code and make comparisons
Close ended
- Easy to analyses and code
- less ambiguous responses, get a comment on responses of interest
- easy to compile.
Define reliability
reliability indicates the extent to which measures are repeatable.
•(i) The consistency of a measuring instrument.
•(ii) The comparison between two parallel measures.
•(iii) the extent to which a re-measurement agrees with the original measure.
•In effect reliability deals with stability and equivalence of the instrument.
Define validity
Validity
•Validity indicates the extent to which the instrument does what it is intended to do.
•
•Internal validity, how the instrument measures the content which is under investigation
•External validity, the property of generalization to the total population.
Name four identity variables in the layout of questionnaire
ie the title of the study, the place of study, the date, the serial number of questionnaire and the name of interviewer, and the house number
Name four requirements in the layout of the questionnaire
Identity variables
Basic characteristics of the respondent: name of the respondent, age, sex, educational status, religion, occupation, marital status, etc
Body of Questionnaire
•Group the questions on the topic
Below are statements involved in the layout for questionnaires Start with fairly general question that can put the respondents at ease. True or false
•Show that the research is interested to learn about the respondent’s views rather than testing him/her. True or false
•Organize questions to flow into one another true or false.
•skip from topic to topic.
–ie exhaust questions on one topic before moving to another. True or false
True
True
True
False
There is a need to have pre-testing of questions. True or false
•In assessing the meanings of question, duration of study in terms of field operation and
•determine which question to drop or add.
True
Explain how interviews are done
Interview
•Selection and training of interviewers for the study and respondents.
- Interviewer, introduce himself/herself.
- Identify eligible respondents and if time is inconvenient, to make new appointment.
•Ask the question exactly as they are written verbatim, avoid giving your own opinion and make the respondent comfortable.
Define single blind randomized trial and double blind trial
Single-blind randomized trial.
–The identity of the allocated treatment is concealed from the patient.
•Double blind trial.
–The doctor, or the technical expert who assesses response also as well as the patient is unaware of the treatment identity. Whether
What is a placebo
medicine or procedure prescribed for the psychological benefit to the patient rather than for any physiological effect.
a substance that has no therapeutic effect, used as a control in testing new drugs
Why is ethics important in sampling
Abuses in the past and some still persist
•Nazi experiments and the Nuremberg trials 1947
•* Tuskegee Syphilis Study – 1932 – 1972
- Jessie Gelsinger – Univ.of Penn. (Died
in a genetic study in 1999)
Name four reasons why the Nuremberg code is good
Voluntary consent essential •Yield fruitful results •Avoid all unnecessary suffering •No predictable death/disability •Risk not greater than importance of problem •Participants can stop anytime •Investigator must stop if harmfu
What is the declaration of helsinki
Based on good science •Reviewed by independent committee •Importance in proportion to risk •Interest of subjects “must always prevail over interests of science and society” •Respect privacy •Informed consent
What are the basic rules in informed consent
Short words, short sentences •Large type fonts, wide margins •Avoid technical language/ jargons •Avoid passive tense •Use language of subject population •Consent forms must be signed and dated by participants/legally authorised representati
As a rule for frequency distribution table the classes should be mutually exclusive. There should be no overlapping classes and classes should be continuous As in if there ar eno values for 21-30 don’t skip from 10-20 and go to 31-40 but do it 20-21,21-30,31-40
True or false
True
The 0.5 is subtracted from the lower class limit and not the upper class limit and the 0,5 is added to the upper class limit true or false
True
Relative frequency is equal to
Frequency divided by total frequency times 100
What is hypothesis testing and how is it done
Decisions process for evaluating a claim about a population which could be true or false
- Define population under study
- State the hypothesis to be investigated
- Give a significant level
- Select a sample from the population
- Collect data
- Perform necessary calculations to perform the test
Symbol for population is mew true or false
True
What is null and alternative hypothesis
Null - there is no difference between the parameters and a specific value or between two values
Alternative: there is a difference
Most statistical tests are done based on null hypo and if the test doesn’t meet the test of significance it means the null hypothesis is true and alternative hypo is rejected
True or false
True
A claim is stated as an alternative hypo true or false
True
variable that interferes with other variables in the study is called. a confounding variable.
True or false
True
How is variable name different from label name
It is shorter and less detailed
Difference between intervals Hs ratio is that w interval there’s a difference between the rankings and there’s no zero as in someone can’t have an IQ of zero but in ratio the zero is important like someone having bp of zero is important true it false
True
Yes of association is the relationship between two variables and if the relationship is significant true or false
True
There are no fractions in discrete data . Weight is and example of ratio cuz you can’t have zero weight and kelvin as a thermodynamic temperature is another example cuz you can’t have zero kelvin . Statistics are used to estimate parameters. For sensitive info snowball is used true or false
True
