SML Previous Year Question Paper
Uploaded by
abcdxyzlmnop3232SML Previous Year Question Paper
Uploaded by
abcdxyzlmnop3232Common questions
Powered by AIRandom sampling is crucial as it ensures that every member of a population has an equal chance of selection, thereby achieving a representative sample. It reduces biases and supports valid generalizations from sample to population. Not using random sampling may lead to selection bias, where certain population sections are over- or under-represented, possibly distorting findings and leading to invalid inferences about the population characteristics .
The Central Limit Theorem (CLT) states that for a sufficiently large sample size, the sampling distribution of the sample mean approaches a normal distribution, regardless of the population's original distribution. This theorem is foundational in statistics because it underpins methodologies like confidence intervals and hypothesis testing, allowing statisticians to apply normal distribution techniques to sample data irrespective of actual population distribution .
The normal distribution is symmetric, characterized by its mean and standard deviation, applicable generally when population parameters are known or sample sizes are large. The student's t-distribution, heavier-tailed, caters to smaller samples when population standard deviation is unknown; as sample size increases, it approximates the normal distribution. It is commonly used in small-sample hypothesis testing and constructing confidence intervals, providing a correction for the additional uncertainty in estimating the population standard deviation .
The correlation coefficient quantifies the strength and direction of a linear relationship between two variables, ranging from -1 to 1. A value of 1 or -1 indicates perfect correlation, while 0 signifies no correlation. Limitations include its sensitivity to outliers, inability to capture non-linear relationships, and the assumption of a linear relationship leading to possible misinterpretations if nonlinearity exists. Therefore, additional exploratory data analyses may be required to confirm linear relationship assumptions .
Measures of central tendency, such as the mean, median, and mode, provide insights into the center of a data set. The mean offers the arithmetic average, representing the overall level if data values are balanced; it is useful for further statistical analysis like variance calculation but can be affected by outliers. The median, the middle value when data is ordered, offers a robust central measure resistant to outliers. It is particularly useful for skewed distributions. The mode identifies the most frequently occurring value, useful in categorical data representation .
The bootstrap method enhances variability estimation by resampling with replacement from the sample data, allowing approximation of the sampling distribution of a statistic without relying on traditional parametric assumptions. This technique enables robust estimation of standard errors and confidence intervals, particularly useful when sample sizes are small or population distributions are unknown. In modern statistical analysis, bootstrap is instrumental in complex models and non-parametric contexts, offering flexibility and insight into estimator variability .
Histograms and kernel density estimates (KDE) both illustrate data distribution yet differ in approach. Histograms do so by grouping data into bins and displaying frequencies, offering ease of interpretation but requiring careful bin selection. This is useful for visualizing discrete data or when a quick understanding of distribution shape is needed. KDE, a non-parametric estimation, smooths data into a continuous probability density curve, allowing for finer detail in highlighting distribution aspects such as modality and peaks without bin influences, facilitating analysis of continuous data properties .
Point estimation involves providing a single value (point) as an estimate of a population parameter, such as using the sample mean to estimate the population mean. It is precise but lacks information about the estimation accuracy. Interval estimation, in contrast, involves providing a range, or interval, within which the parameter likely falls. This accounts for sample variability and offers a confidence level indicating the probability the interval contains the parameter. Interval estimates address the limitations of point estimates by providing a reliability context .
Boxplots offer a summary of the data distribution in five-number format (minimum, Q1, median, Q3, maximum), highlighting data spread, symmetry, and potential outliers efficiently. Unlike histograms, which can obscure detail by aggregation and binning choices, boxplots provide a direct visual cue for comparing multiple data sets without overlaying distributions. For example, in comparing test scores between classes, boxplots can succinctly display medians, IQRs, and outliers for both classes, facilitating clear comparative analysis .
The interquartile range (IQR) measures statistical dispersion, representing the range within which the central 50% of data lies, specifically between the first (Q1) and third quartiles (Q3). It is crucial for identifying outliers, as data points lying below Q1 - 1.5*IQR or above Q3 + 1.5*IQR are often considered outliers. This method leverages the robustness of the median, as IQR is stable under transformations and not affected by extreme values .