Research Hypothesis - Assignment Example

Assignment number: of Convenors: Due week 7 , Friday 23rd of April, Semester Before 5:00pm. Part 1 (STUDY TWO) (1) Provide research hypothesis for the study based on the given study aim (Mark 2) The null and alternate hypotheses are H0: There is no association between the duration of breastfeeding and the risk of asthma later in life in Finland children. H1: There is an association between the duration of breastfeeding and the risk of asthma later in life in Finland children. (2) Which design is appropriate in investigating the question in STUDY TWO; provide a rationale as to why you have made this choice? (Note –there may be more than one suitable choice) (Mark 4) A population-based cohort (correlational) study design is appropriate in investigating the question in STUDY TWO. A rationale as to why I have made this choice is that the study is observational, unit of study is individual, and time is prospective (two-time points). (3) Using the most appropriate design provide a plan of your study design (Mark 3) The study will be conducted on randomly selected urban-suburban municipality from city of Finland. The sample will be all children born between January 1, 2003, and December 31, 2008. The baseline study population will be children between the ages of one to seven years. Two-time points I will collect data (6 to 7 years cohort). There will be only one group. (4) What are the names of variables, are there more than two variables in this study? How might you measure these? (Mark 4) The first questionnaire will consist of data related to age, gender, any allergic disease, single parent/guardian, highest level of parental education, hairy/feathery pets, type of day care, etc at the first time. The second questionnaire at the end of the follow-up period will includes all the questions from first questionnaire besides data related to duration of breastfeeding (independent variable) and current asthma (dependent variable). The duration of breastfeeding will be divided into categories as 0-3 months, 4-6 months, 7-9 months, 10-12 months, and more than 12 months. The current asthma will be measured using Yes/No response. The survey method will be Researcher-administered. (5) Who will be your study sample? What is your target population and what is sampling frame? What are your comparison groups? Are they groups and/or time points? (Mark 4). The sample will be all children born between January 1, 2003, and December 31, 2008 in urban-suburban municipality from city of Finland. The target population will be all Finland children. The sampling frame will be the Central Population Registry from which the initial data for the baseline study population such as address, phone number will be collected for all children born between January 1, 2003, and December 31, 2008. There will be only one group and two time points as described earlier in (4). (6) Can your dependant variable/s be directly attributed to the independent variable/s? Are there potential sources of bias or error related to the design choice? What might they be? (Mark 4) Yes, I think that there is an association between the duration of breastfeeding and the risk of asthma later in life. Therefore, the risk of asthma later in life depends upon the duration of breastfeeding. There are potential sources of bias or error related to the design choice such as Selection bias (recruitment and response) - medium or high, Information bias - medium, losses to follow up, and confounding factors. PART 2 STUDY ONE (Open Assignment 1.1.sav dataset) Two research questions are: 1. What are the demographic characteristics of cancer patients? 2. Is years of smoking, number of cigarettes smoked and alcohol consumption in cancer patients different to control group? 1. Hypothesis: Write-up the scientific hypotheses that you want to test to address the research question 2. (1 Marks) The three null and alternate hypotheses are H0: Years of smoking is similar in cancer patients and control group. H1: Years of smoking is different in cancer patients and control group. H0: Number of cigarettes smoked is similar in cancer patients and control group. H1: Number of cigarettes smoked is different in cancer patients and control group. H0: Alcohol consumption is similar in cancer patients and control group. H1: Alcohol consumption is different in cancer patients and control group. 2. Variables: For each research question identify- Variables/measurement type for each variable (Mark 2) Variable Measurement Type (Scale) Sex Dichotomous (or nominal) Age Interval Age at diagnosis Interval Living environment Nominal Living condition Nominal Contact with X-Ray at work Dichotomous Contact with radioactive substances at work Dichotomous Work in air pollution Dichotomous Further education Dichotomous Socio-economic status Ordinal Years of smoking (years) Ordinal Number of cigarettes smoked per day Interval Alcohol consumption Dichotomous 3. Data management and data cleaning (1 mark) Variable Missing Data Values Case No missing values Sex No missing values Age blank Age at diagnosis -1, blank Living environment blank Living condition blank Contact with X-Ray at work blank Contact with radioactive substances at work blank Work in air pollution blank Further education blank Socio-economic status blank Years of smoking (years) No missing values Number of cigarettes smoked per day blank Alcohol consumption blank Four duplicate cases (PatientID) were detected and deleted using Data Identify Duplicate Cases as below: 94027741.2 94027741.2 (Deleted) 96015990.5 (Deleted) 96015990.5 96066763.5 (Deleted) 96066763.5 97061147.1 (Deleted) 97061147.1 4. Univariate Analysis: For each variable, choose the appropriate numerical statistical summery and or graphical summary for cancer patients group, provide the results of statistical assumption if the variables are interval or ratio data type (3 Marks) The Statistical Assumption (normality) Check for Interval and Ratio Data (Question 1) Statistics AGE AGE AT DIAGNOSIS N Valid 116 116 Missing 0 0 Mean 43.92 38.78 Median 44.50 40.00 Std. Deviation 5.819 5.731 Skewness -.578 -1.089 Std. Error of Skewness .225 .225 Kurtosis -.273 .822 Std. Error of Kurtosis .446 .446 Minimum 30 20 Maximum 54 45 Age Mean is within 10% of the median. Calculating mean ± 3SD = 43.92 ± 3 × 5.82 = 26.46 min, 61.38 max. Min is 26.46, max is 61.38, which are close to 30 and 54. The skewness and kurtosis are well within the range of -3 and +3. The assumption of normality is accepted. Age at diagnosis Mean is within 10% of the median. Calculating mean ± 3SD = 38.78 ± 3 × 5.73 = 21.59 min, 55.97 max. Min is 21.59, max is 55.97, which are close to 20 and 45. The skewness and kurtosis are well within the range of -3 and +3. The assumption of normality is accepted. Therefore, Mean and SD will be used to summarise the variables Age and Age at diagnosis. Table 1 – Demographic variables Variable N Mean (SD) Age 116 43.92 (5.82) Age at diagnosis 116 38.78 (5.73) Table 2 – Demographic variables Variable N (%) Sex Male 65 (56.0) Female 51 (44.0) Living environment City 51 (44.7) Suburbs 46 (40.4) Rural 17 (14.9) Living condition Alone 20 (17.4) With partner 84 (73.0) With parents 3 (2.6) Other 8 (7.0) Contact with X-Ray at work No 108 (93.9) Yes 6 (5.2) Don’t know 1 (0.9) Contact with radioactive substances at work No 107 (93.0) Yes 1 (0.9) Don’t know 7 (6.1) Work in air pollution No 81 (70.4) Yes 15 (13.0) Don’t know 19 (16.5) Further education No 59 (50.9) Yes 57 (49.1) Socio-economic status I&II 47 (40.5) III 45 (38.8) IV&V 13 (11.2) Non-class 11 (9.5) Table 1 and 2 show the demographic characteristics of cancer patients. The average age of cancer patients was about 43.92 years (SD = 5.82). The average age at diagnosis of cancer patients was about 38.78 years (SD = 5.73). Majority of the cancer patients were male (56%). About 44.7% of cancer patients are from city, 40.4% from suburbs, and 14.9% from rural environment. Majority (73.0%) of cancer patients are living with partner. Most (93.9%) of cancer patients have no contact with X-Ray at work. Most (93.0%) of cancer patients have no contact with radioactive substances at work. Majority (70.4%) of cancer patients are not working in air pollution. There is an even response for further education. The Socio-economic status of cancer patients are I&II (40.5%), III (38.8%), IV&V (11.2%), and non-class (9.5%). 5. Bivariate Analysis: Present table (numerical) would you use to summarise the difference between case and control group in years of smoking, number of cigarettes smoked and alcohol consumption once the data are analysed (3 Marks) Mean is not within 10% of the median for control and case groups. Therefore, Median and Range (Min, Max) will be used to summarise the variable number of cigarettes smoked per day for control and case group. Table 3 – Summary statistics of number of cigarettes smoked per day N Median Range (Min, Max) Control (normal) 200 5 50 (0, 50) Case (Cancer patients) 116 10 60 (0, 60) Table 3 shows the summary statistics of number of cigarettes smoked per day. The average (median) number of cigarettes smoked per day for cancer patients is higher (double) as compared to control group (10 vs. 5). The spread (range) of number of cigarettes smoked per day for cancer patients is wider as compared to control group (60 vs. 50). Table 4 – Summary statistics of years of smoking and alcohol consumption Case or Control Variable Control (normal) Case (Cancer patients) N (%) N (%) Smoking in years Less than 1 68 (33.5) 33 (28.4) 1 - 10 37 (18.2) 15 (12.9) 11 - 20 44 (21.7) 17 (14.7) 21+ 54 (26.6) 51 (44.0) Alcohol consumption No 21 (10.4) 16 (13.8) Yes 181 (89.6) 100 (86.2) Table 4 shows the summary statistics of years of smoking and alcohol consumption for case and control group. More numbers of cancer patients are smoking for 21 and more years as compared to control group (44.0% vs. 26.6%). SPSS OUTPUT: Part 2- Study One (Assignment 1.1) Statistics CASE OR CONTROL smoking in years No of Cigarettes Smoked PER DAY CONTROL_normal N Valid 203 200 Missing 0 3 Mean 1.41 9.08 Median 1.00 5.00 Std. Deviation 1.205 10.951 Skewness .085 1.189 Std. Error of Skewness .171 .172 Kurtosis -1.545 .963 Std. Error of Kurtosis .340 .342 Range 3 50 Minimum 0 0 Maximum 3 50 CASE_Cancer patients N Valid 116 116 Missing 0 0 Mean 1.74 12.08 Median 2.00 10.00 Std. Deviation 1.286 13.621 Skewness -.326 1.402 Std. Error of Skewness .225 .225 Kurtosis -1.620 2.230 Std. Error of Kurtosis .446 .446 Range 3 60 Minimum 0 0 Maximum 3 60 smoking in years CASE OR CONTROL Frequency Percent Valid Percent Cumulative Percent CONTROL_normal Valid less than 1 68 33.5 33.5 33.5 1-10 37 18.2 18.2 51.7 11-20 44 21.7 21.7 73.4 21+ 54 26.6 26.6 100.0 Total 203 100.0 100.0 CASE_Cancer patients Valid less than 1 33 28.4 28.4 28.4 1-10 15 12.9 12.9 41.4 11-20 17 14.7 14.7 56.0 21+ 51 44.0 44.0 100.0 Total 116 100.0 100.0 CONSUME ALCOHOL CASE OR CONTROL Frequency Percent Valid Percent Cumulative Percent CONTROL_normal Valid NO 21 10.3 10.4 10.4 YES 181 89.2 89.6 100.0 Total 202 99.5 100.0 Missing System 1 .5 Total 203 100.0 CASE_Cancer patients Valid NO 16 13.8 13.8 13.8 YES 100 86.2 86.2 100.0 Total 116 100.0 100.0 Missing System Total STUDY TWO (Open Assignment 1.2.sav dataset) Two research questions: 1. Are there differences between three birthweight groups in maternal lead level? 2. Is child’s birth weight associated with maternal age of the mother (at first child birth)? 1. What are the variables involved in answering the research question/s (2 marks) Research Question 1 Two variables involved are birth-weight groups (Independent variable), and maternal blood lead levels (Dependent variable). Research Question 2 Two variables involved are maternal age of the mother at first child (Independent variable), and child’s full term birth weight (Dependent variable). 2. What is their level of measurement (type); (1 mark) Variable Measurement Type (Scale) Full term birth weight (in grams) Ratio Birth weight by group Ordinal Maternal age of mother (in years) Interval Maternal blood lead levels (micrograms per decilitre) Ratio 3. Write-up the scientific hypotheses that you want to test to address the research question/s; (1 mark) Research Question 1 The null and alternate hypotheses are H0: There are no differences between three birth-weight groups in maternal lead level. H1: There are differences between three birth-weight groups in maternal lead level. Research Question 2 The null and alternate hypotheses are H0: Child’s birth weight is associated with maternal age of the mother (at first child birth). H1: Child’s birth weight is not associated with maternal age of the mother (at first child birth). 4. What table (numerical) or graphs would you use to summarise the data? Provide tables or graphs with appropriate labels; (2 marks) For research question 1, a boxplot is an appropriate graph and for research question 2, a scatterplot is an appropriate graph. Figure 1: Boxplot of maternal blood lead level by three birth-weight groups Figure 2: Scatterplot of full-term birth weight against maternal age of mother 5. Provide a list of assumptions that will need to be met to apply the test(s) validly; (1 mark) A list of assumptions that will need to be met to apply the test(s) validly for both research questions are given below: Research Question 1 Independence of observational units Normality of dependent variables Homogeneity of variances across comparison groups Research Question 2 Independence of observational units Normality of dependent variables The Statistical Assumption (normality) Check for Dependent Variable (Question 1) Statistics maternal blood lead levels - micrograms per deciliter N Valid 250 Missing 0 Mean 4.1184 Median 3.5794 Std. Deviation 2.23943 Skewness 3.254 Std. Error of Skewness .154 Kurtosis 22.402 Std. Error of Kurtosis .307 Range 22.25 Minimum 1.24 Maximum 23.49 Maternal blood lead levels Mean is not within 10% of the median. The skewness and kurtosis are not within the range of -3 and +3. Histogram is not bell-shaped. The assumption of normality is not accepted. Therefore, Median and Range (Min, Max) will be used to summarise the dependent variable maternal blood lead levels. Table 1 – Summary statistics of the dependent variable maternal blood lead levels N Median Range (Min, Max) Maternal blood lead levels 250 3.58 22.25 (1.24, 23.49) The Statistical Assumption (normality) Check for Dependent Variable (Question 2) Statistics full term birth weight in grams N Valid 250 Missing 0 Mean 3204.47 Median 3178.00 Std. Deviation 620.986 Skewness .162 Std. Error of Skewness .154 Kurtosis -.171 Std. Error of Kurtosis .307 Range 3316 Minimum 1709 Maximum 5025 Full term birth weight Mean is within 10% of the median. Calculating mean ± 3SD = 3204 ± 3 × 621 = 1341 min, 5067 max. Min is 1341, max is 5067, which are close to 1709 and 5025. The skewness and kurtosis are well within the range of -3 and +3. Histogram is bell-shaped. The assumption of normality is accepted. Therefore, Mean and SD will be used to summarise the dependent variable full term birth weight. Table 2 – Summary statistics of the dependent variable full term birth weight Variable N Mean (SD) Full term birth weight 250 3204.47 (620.99) 6. Describe results for both research question 1 and research question 2. (2 marks) Research Question 1 As shown in figure 1, there appear differences between three birth-weight groups in maternal blood lead level. Therefore, average maternal blood lead level differs by three birth-weight groups. In addition, there appears substantial maternal blood lead level variation within each birth-weight. Research Question 2 As shown in figure 2, there appears a positive linear relationship between full term birth weight and maternal age of mother. Therefore, child’s birth weight is associated with maternal age of the mother (at first child birth). SPSS OUTPUT: Part 2- Study Two (Assignment 1.2) Statistics maternal blood lead levels - micrograms per deciliter Low birthweight N Valid 33 Missing 0 Mean 4.0167 Median 3.5086 Std. Deviation 2.40766 Skewness 1.134 Std. Error of Skewness .409 Kurtosis 1.644 Std. Error of Kurtosis .798 Range 10.44 Minimum 1.24 Maximum 11.68 Lower end of normal birthweight N Valid 60 Missing 0 Mean 3.4214 Median 2.9523 Std. Deviation 1.55170 Skewness .828 Std. Error of Skewness .309 Kurtosis -.351 Std. Error of Kurtosis .608 Range 5.79 Minimum 1.39 Maximum 7.18 Normal birthweight N Valid 157 Missing 0 Mean 4.4062 Median 3.7572 Std. Deviation 2.37333 Skewness 3.879 Std. Error of Skewness .194 Kurtosis 26.516 Std. Error of Kurtosis .385 Range 21.62 Minimum 1.87 Maximum 23.49 Correlations maternal age of mother full term birth weight in grams maternal age of mother Pearson Correlation 1 .180(**) Sig. (2-tailed) .004 N 250 250 full term birth weight in grams Pearson Correlation .180(**) 1 Sig. (2-tailed) .004 N 250 250 ** Correlation is significant at the 0.01 level (2-tailed). Read More