Purpose: This study assessed the accuracy of linear and angular measurements on panoramic radiographs taken at different positions in vitro. Materials and Methods: Two acrylic models were fabricated from a cast with normal occlusion. Straight and $75^{circ}$ mesially and lingually angulated pins were placed, and standardized panoramic radiographs were taken at standard position, at an $8^{circ}$ downward tilt of the occlusal plane compared to the standard position, at an $8^{circ}$ upward tilt of the anterior occlusal plane, and at a $10^{circ}$ downward tilt of the right and left sides of the model. On the radiographs, the length of the pins above (crown) and below (root) the occlusal plane, total pin length, crown-to-root ratio, and angulation of pins relative to the occlusal plane were calculated. The data were subjected to repeated measures ANOVA and LSD multiple comparisons tests. Results: Significant differences were noted between the radiographic measurements and true values in different positions on both models with linear (P<0.001) and those with angulated pins (P<0.005). No statistically significant differences were observed between the angular measurements and baselines of the natural head posture at different positions for the linear and angulated pins. Conclusion: Angular measurements on panoramic radiographs were sufficiently accurate and changes in the position of the occlusal plane equal to or less than $10^{circ}$ had no significant effect on them. Some variations could exist in the pin positioning (head positioning), and they were tolerable while taking panoramic radiographs. Linear measurements showed the least errors in the standard position and $8^{circ}$ upward tilt of the anterior part of the occlusal plane compared to other positions.