DATA RELIABILITY AND VALIDITY

The data are of two kinds, the subjective data obtained from the questionnaires and the objective data from the physical site measurement. Here, the reliability and validity of the data will be discussed in two sections. The first will discuss the subjective data and the second will discuss the objective data.

4 . 8 . 1

Reliability And Validity Of The Subjective Data

The reliability of the subjective data that will be gathered in this research is preassessed using the test-retest method in the EPA (Energy Passive Architecture) study. The reliability is gauged by repeating the test on the same person using the same method. In the EPA study the questionnaire was distributed thrice; once at the start of the study, then 12 months later, and then again at 24 months. The intention of carrying this out was to check if the way the questions were phrased gave rise to different answers as occupants maybe "infatuated" by the newness of the building and if occupants appreciation of the building changes. The EPA study test-retest using the same measure showed that generally the score on the scale moved half a point up or down and statistically there were little difference in the distribution. This movement is related to the occupants settling in the building rather than the accuracy of the judgement. From this, it was found that occupants appraisal after 12 and 24 months was close. Therefore one of the criteria of the implementation of the questionnaire was that occupants must have occupied the building more than 12 months. The correlation coefficient from the test-retest method gives measure of stability because they relate to consistency over time (The new questions added on did not have the test-retest method).

To further enhance the reliability of the subjective data, the alternate form method is employed in the questionnaire. This is achieved by using other questions that address similar aspects but are differently phrased in the same questionnaire. The test carried out on the same questionnaire automatically ensures that the measurements are repeated

under the same conditions'^. Testing on the same questionnaire also takes care of different events^^ occurring between the two tests which may influence their views.

The subjective data is also validated by flipping the assessments in the questions to ascertain that the respondents are aware and had thought carefully whilst answering the questionnaire and not ticking the questionnaire at random.

To ascertain that the scale is valid, it is necessary to ensure that it is measuring what it is designed to measure so that difference between individuals' scores can be taken as representing true differences in the characteristic under study. The subjective data obtained from the methodology is valid for the following reasons:

1. It has cumulative validity - The success of the questionnaire measuring what it sets out to measure is supported by other studies already using similar methods such as BUS. Besides the semantic scale used in it is proven to have substantial validity as a general measure of perception.

2. It has ecological validity as the study is carried out in a natural environment 3. It's representativeness of the office occupants due to the big sample size.

4. Researcher effect on the normal condition is minimised by using an unobtrusive method.

5. It has theoretical (face) validity^ ^ as the questions refer to both symptom occurrence and occupants appraisal of the appearance from the work position.

The practical difficulty o f the this test-retest method is however self-evident; if a person was submitted to repeat questioning, a comparison o f the two sets o f results would hardly serve as an exact test o f reliability, since they could not be regarded as independent. As occupants may remember their first answer and give a consistent retest answer, an action which would make the test appear more reliable than is truly the case. The tactic for overcoming this was to make the questionnaire extensive enough. Where different events do occur to influence assessm ent the scores are not comparable, so the difference between the two is a mixture o f reliability and change in the characteristic itself; the effect o f this is that an underestimate o f reliability is obtained.

^ ^ The measure is said to have theoretical validity when the finding complies with the theoretical finding o f the decipline i.e they do not contradict already existing rules.Theoretical validation are o f 3 types: 1. Face validity - seems to measure what it is expected to measure. Findings comply with the

theoretical finding o f the decipline i.e they do not contradict already existing rules.

2. Content validity - when it covers all aspects o f research topic. Elements that are considered to be important aspects associated with appearance appraisal.

3. Construct validity - when a theoretical construct is valid, whereby the researcher assumes to be true the type and degree o f association between the scale and other variables based on the theoretical consideration and then examines these associations to see whether they confirm to the expectation (the essence o f construct validity is its dependence on theory and the examination o f the observed association is as much a test o f the theory as o f the scale o f validity). The weakness o f this check is the relation between subjective assessment and personal preference. A positive result would give increased confidence in the scale. The existence o f good correlation acts as supporting evidence. Where correlation is low, conformity o f these correlations to expectations provides the necessary test (M oser and Kalton, 1971). Better appearance o f work position propogates fewer symptom occurrences is the research topic. Comparing the group with good appearance and bad appearance (two groups known to have different attitudes to the research issue) satisfies the construct validity and thus the theoritical validity.

It has communicative validity - by administrating additional questioning to building occupants. LINK required the occupants that are willing to further participate in the study to keep a diary.

The sensibility of the subjective assessments are also established in the factor analysis whereby the assessments do not contradict one another. Thus, a work position that was appraised as pleasant would also constantly be assessed with positive remarks.

4 . 8 . 2

Reliability And Validity Of The Objective Data

The objective data from the physical site measurement is divided into two parts. The first part is that of the geometric type that describes space and object dimensions, and the second part describes the light parameters. The reliability of both sets of measurements are closely associated with the instruments that will be used (refer to section 4.3.2-ll(ii)). The instruments themselves are calibrated in laboratories. By satisfying the criteria set for the instrument selection, the data is automatically made valid and reliable. From here on the geometric data of the visual environment are straight forward. However, the light parameters are rather complicated and vary dynamically. The appropriateness of the instrument alone is insufficient to verify these data. These data need to be further verified (giving the right reading) by repeating the measurements, at least twice. The measurement is said to be reliable as the repeat measurements made under the same conditions gives the same reading.

The light parameter data are obtained under overcast conditions and valid only for this condition. Stating this, a spot reading is representative of the daily variation. This daily representation of the data is also verified by the 2hrs interval measurements that is carried out throughout the working day. Readings showed that they are relatively constant throughout the day. Thus, the physical light measurements represented the daily variation but not the yearly variation as sunny days and sunlight penetration was not taken into account. However, as overcast days represent 84% of the year, and we could say that the light parameters that are measured represent the condition for 84% of the year, that is when the external illuminance is less or equal to 10,000 lux.

4 .9

CHRONOLOGICAL ORDER OF THE METHODOLOGY

1. Select building

2. Primary LINK assessment 3. Distribute questionnaire 4. Collect questionnaire

5. Log in survey questionnaire data

6. Conduct factor analysis to Identify Cluster

7. Go into building to do objective physical measurements (simultaneously with the LINK team)

8. Distribute Eye health questionnaire

9. Log in objective physical measurements data and Eye health questionnaire data

Where organisations are unwilling to allow this, the investigation becomes impossible and will be abandoned as there is no way to link the physical data to that of subjective assessment nor the occurrence of eye symptoms^

This methodology will be tested in a pilot study which will be discussed in the next chapter.

^ Although the inability to correlate physical measurements to the questionnaire renders much o f the study obsolete, in these cases part o f the study could still be carried out, that is the relation o f occupants perception to the occurrence o f eye symptoms. This is because when correlating occupants perception to the occurrence o f the symptoms no work on position identification is required as data are retrieved from one questionnaire. However, the analysis is only half done as it could not relate to the physical attribute o f the environment.

Chapter 5

TESTING OF THE METHODOLOGY: The Pilot Study

5 .0

INTRODUCTION

This chapter describes the testing of the methodology that has been developed. It focuses on South Lakeland District Council in Kendal as a pilot study to test out the methodology.

The chapter consists of 8 sections:

1. Will give a brief description of the building, explain the choice of building. 2. Will describe the building in detail. The description of the building includes;

i. the spatial organisation, ii. the room surfaces,

iii. the electric lighting installation, iv. the window and daylighting,

V the light pattern On the VWS,

vi the visual environment in the office. 3. Relates the collection of the data. This includes;

i. the schedule when the pilot study was conducted, ii. the floors used in the study,

iii. the problems that were encountered on site were discussed. 4. Relates the data.

5. Relates the logic structure for the analysis. 6. Relates the analysis procedure.

6. Relates the results of the analysis.

7. Discusses the problem with the methodology adopted and the amendments made.